Professional Documents
Culture Documents
Volume 5
Edited
by Emil A. Fellmann
Norbert Wiener
1894-1964
by P. R. Masani
1990
Birkhauser Verlag
Basel Boston Berlin
Author's Address:
P.R. Masani
Department of Mathematics
University of Pittsburgh
Pittsburgh, Pennsylvania 15260
USA
Contents
12
13
Prologue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
29
34
39
45
Contents
66
The young war enthusiast. Futile efforts to enlist in the armed forces.
Staff work with the Encyclopedia Americana. Computer at the U. S.
Army Proving Grounds. Private in the U. S. Infantry. Fowlersque
touch.
7
71
Slow promotion and resulting competitiveness. Mal-allocation of national income. Margaret Engemann and marriage. Search for positions
abroad.
9
132
A) Radiative equilibrium in the stars. B) Significance of the HopfWiener equation; causality and analyticity. C) Paley and the Fourier
transformation in the complex domain.
Contents
12
139
160
181
A) Project D.!. C. 5980: Its military and scientific significance. B) Idealization of the flight trajectory and resulting tasks. C) The mathematical
theory. D) Operation of the anti-aircraft predictor. E) Resolution of the
man-machine concatenation. F) The problem of transients. G) Secrecy,
overwork and tension. H) Kolmogorov's paper and prediction theory.
I) The Kolmogorov-Wiener concatenation. Note 1: Bigelow'S design.
Note 2: The equation 14C (3); theory of instrumentation.
15
197
218
10
Contents
17
239
18
Cybernetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
251
19
272
A) Automatization and the labor movement. B) The educational challenge. C) Anti-homeostatic aspects of the American economic system.
D) Wiener's economic ideas. E) The control of conflict in civil society.
F) Communal information. G) The grammar of the social sciences.
H) Wiener and Marxism.
20
295
A) Wiener's thoughts on geopolitics. B) Wiener's letter to Walter Reuther on geopolitics, and von Clausewitz's principles on war. C) Typeclassification in military science. D) The pitfalls in computerized atomic
war games. Time scales. E) Synopsis of Wiener's military thought.
F) Von Neumann's position; science and human welfare. G) The Black
Mass, twentieth century. Note: Von Clausewitz's principles of war.
21
318
Contents
11
long-time State. The mandate from heaven. F) Sin grows with doing
good: Tragedy and catharsis. G) Rational Logos, or rational and
altruistic Logos?
22
335
349
A) The lovable quirk. B) Wiener as public figure and teacher. C) Resignation from the National Academy of Sciences; muddled ideas on
science academies. D) Unobjective attitude towards Harvard. E) Wiener
and George David Birkhoff.
24
Epilogue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
366
371
375
376
377
391
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
392
400
Subject Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
407
415
12
13
Acknowledgment
This book is an enlargement of an article in the Encyclopedia of Computer Science and Technologyl, that has grown and grown as I became
more and more aware of the enormity and profundity of Wiener's
thought. Its scope and purpose are indicated in the Prologue (Chap. J,
A,B,C), and only the pleasant task of making acknowledgment remams.
Although the manuscript was seen by a few publishers, it might
never have seen the light of day had Professor P.L. Butzer of the
Technische Hochschule in Aachen not passed on a copy to Dr. E. A.
Fellmann of the Euler Archiv in Basel. My sincere thanks go to him,
and to Dr. Fellmann for his invitation to publish it in the series Vita
Mathematica.
Let me rethank Professors J. Belzer and A. Kent and Professors
W. T. Martin, G. Birkhoff and M. H. Stone (deceased) for their initial
help in 1979-80. Second, I want to thank Ms. Helen Samuels of the
Archive Section at MIT for placing Wiener's unpublished material for
my use in May 1981, and Professor I. E. Segal of MIT for permission
to use his list of Wiener's doctoral students. I am grateful to Mr. E. F.
Beschler of Birkhauser Boston, Ms. K. McLaughlin and Professor R. 1.
Parikh for reading and criticizing the manuscript in whole or part, to
Professor 1. Benedetto for providing me with inaccessible biographical
data, and to my friend Mr. K. N. Karanjia of Bombay for sending me
a less-known paper .of Dirac from which I have quoted. Mr. Beschler
was also very helpful in settling some copyright difficulties.
The following individuals were very generous and forthright in
sending me photographs for the book: Mr. 1. Bigelow, Ms. A. Bogovich,
Dr. F. Perrin, Professor N.N. Rao, the late Mrs. Virginia Rosenblueth,
Dr. W.A. Stanley, Dr. K. Thompson, Mr. T. Walley Williams and Mrs.
Margaret Wiener. My sincere thanks go to them, and to Mr. M. Roper
of our Fine Arts Department for his fine reproductions of some of the
other photos. I also want to thank M r. M. Yeates of the MIT Museum
Vol. 14 (1980), pp.72-136, Editors J. Belzer, H. Holtzman & A. Kent,
Marcel Dekker, New York.
14
Acknowledgment
P.R. Masani
15
Prologue
16
1 Prologue
automatic factory, but also on the ontogenetic and phylogenetic learning of biological species. He showed how communication is the dynamic
cement that sustains a community, and how sponsored misinformation
is a danger. In his socio-economic writings he suggested ways to relieve
the non-homeostatic aspects of the economy. On military matters he
brought to bear a unique perspective based on his understanding of
typology, and time scales of interacting mechanisms. This great thinker
also reflected on the problem of war, and on the wider problems of evil
and of human governance, shedding new light on old theological issues.
In addition, Wiener partook of humane endeavor. He conferred
with the labor leader Walter Reuther on the circumvention of workingclass adversity during the automatization of the United States economy.
His correspondance reveals a feeling concern for suffering humanity,
and a strong sense of social justice and social responsibility.
Although Wiener wrote more than twenty years ago, his writings
bear vitally on our present predicament. Indeed, much contemporary
analysis and commentary seems banal and naive when set beside the
profundity of his thought. Had this thought received better attention
forty years ago, we would not be facing so many acute problems today.
We would be spared from even more acute problems in the future, were
we to heed his thought today.
B The difficulties of a biographer
17
1B Difficulties of a biographer
Charles Sanders
Peirce, 1839-1914.
Courtesy of
Dr. WilliamA. Stanley,
National Oceanic and
Atmospheric
Administration,
Washington, D.C.
18
1 Prologue
When people asked where to find his house in Gottingen he would say quite
naively, "You will have no difficulty finding it. It is the finest house in the
town". [56g, p. 24]
19
No evidence is provided for what "seems" and what "probably" happened. Next comes the conjecture: Newton "was able to channel his
intellectual and psychic energies with such intensity and even ferocity on
scientific discovery", "may be even" because of "his psychological handicaps" {B9, p. 133}. All this is quite dubious, and sheds no clear light
on Newton's creativity.
The same author continues: "He (Newton) spent at least as much
time, perhaps more, on alchemy and Biblical chronology as on physics"
{B9, p. 123}. Aside from the dubiety of the time-estimation, what is
lacking is allusion to the fact that Newton was not unique in this regard,
that Kepler dealt with astrology, Leibniz spent considerable time on
princely genealogy, and Pascal was both scientist and mystic and be-
20
Prologue
21
Consider, for instance, the triviality that the product of the sum of two
numbers with a third is equal to the sum of the product of the first and
third and the product of the second and third. Mathematical symbolism
allows us to dispose of this 32-word triviality by the simple formula
(a
b) . c = a' c
+ b c.
(1)
Gone is the long-winded verbosity, the hallmark of bureaucratic chicanery and fake labor; instead, in Whitehead's words,
... we can make transitions in reasoning almost mechanically by the eye,
which otherwise would call into play the higher facilities of the brain. {W7,
p.6I}
This little book of Whitehead for the beginner provides an excellent insight
into the spirit and scope of mathematics.
22
1 Prologue
Thus from the equality (1), it takes but a glance to see, for instance, that
(a
b)'c - bc
ac.
Only mathematical symbolism has this power of enlisting both mind and
eye in logical reasoning. The eye is able to discern new and interesting
results, which would otherwise go undetected. Between 500 B. C. and
1850 A.D., mathematics made great strides, but the science oflogic was
dormant. This was due in large measure to its bondage to prose. Its great
advances began in the 1850s when G. Boole and C. S. Peirce so symbolized the subject that algebraic operations similar to those for numbers could be performed, and the eye and mind could work together.
At the foundational level of a science, where it is easy to commit
logical errors, the rigidity of mathematical symbolism plays a restraining
role. B. Russell, one of Wiener's most important mentors, has explained
how this happens:
Now, in the beginnings, everything is self-evident; and it is very hard to see
whether one self-evident proposition follows from another or not. Obviousness is always the enemy to correctness. Hence we invent some new and
difficult symbolism, in which nothing seems obvious. Then we set up certain
rules for operating on the symbols, and the whole thing becomes mechanical.
In this way we find out what must be taken as premiss and what can be
demonstrated or defined. For instance, the whole of Arithmetic and Algebra
has been shown to require three indefinable notions and five indemonstrable
propositions. But without a symbolism it would have been very hard to find
this out.{R6, p. 77)
J(P1dx 1 + P 2dx 2
c
P 3dx 3),
(2)
23
curl P
(OP 3
oX 2
oP3).
oX2
SP' dx.
(3)
The aesthetic aspect of mathematical formalism gives deep insights into the structure of the universe, and has therefore considerable
heuristic value in scientific discovery. Of the many examples of this, one
of the most inspiring is Maxwell's discovery of the electromagnetic
theory of light. It involves the concept of the curl just mentioned.
The state attained by electromagnetic theory in the late 18508
can be summed up in four equations governing the electric and magnetic
field vectors4 E and H. Only two of these concern us, viz.
oH
ot
c' curl E
' o=
curl H.
(4)
3 Thus on the right hand side of (2) the subscripts 1,2, 3 are simply arranged
one after the other. On the left hand side the subscripts are mixed-up, but
according to a pretty pattern. Thus, in the coefficient of t " not I but 2 and
3 appear; in the coefficients of t 2, not 2 but 3 and I appear; and similarly for
the coefficients of t 3' Also, in each coefficient the two subscripts appear once
in the numerator and once in the denominator. Finally, notice that if we take
1,2,3 cyclically, i. e. think of I as succeeding 3, then the terms in coefficients
having the minus sign are just those in which the subscript in the denominator succeeds that in the numerator.
4 The reader should think of a vector as a quantity possessing both a magnitude and a direction. An example is the gravitationalforce that the sun exerts
on the earth. This has a definite direction at each moment, viz. that from the
earth's center to the sun's center. But it also has a strength or magnitude
depending on the distance between the centers. Much the same is true for the
electric and magnetic forces surrounding a body. Another example of a
vector is the displacement, such as that of a billiard ball on a table.
24
Prologue
oH
ot
ccurlE
oE
' ot
-
c curl H.
(5)
The second equation in (5) flew in the face of evidence, for it conflicted
with the second equation in (4). But this did not deter Maxwell. He had
a strong intuitive feeling that the mathematically more beautiful pair (5)
had to be right.5 Assuming the equations (5), he proceeded to find out
what they entailed. The elimination of H from the two equations in (5)
yields the equation
(6)
Notice the complete symmetry (but for the minus sign) in the pair of
equations (5). According to (5), E depends on the fluctuations ofH in exactly
the same way as H depends on the fluctuations of E. In (4) there is no such
symmetry in the roles of E and H. If we accept, as Hardy did, that "there is
no permanent place in the world for ugly mathematics", then we must place
our bet on the equations (5).
25
physicist from Kepler on. It is based on the faith that the universe is
fundamentally harmonious, and that therefore the enunciation of its
fundamental principles will be symbolically aesthethic. It is the faith
which Einstein expressed in the words, "I believe in Spinoza's God who
reveals himself in the harmony of what exists".
The large-scale production of radio waves, which Hertz had
found and whose existence was predicted by Maxwell nine years earlier
on the basis of mathematical harmony, has changed human technology
in profound ways that has affected us all. Electronics has become a very
important branch of engineering. These changes were stepping stones to
the further changes that have brought on our age of automatization. In
these later transitions too, the quest for mathematical harmony has
played a vital role, as we shall see from Wiener's work. We would refer
the reader who seeks to understand the place of mathematical harmony
in the life of man to Whitehead's two essays, Mathematics as an element
in the history of thought {W8, Ch. II} and Mathematics and the Good
{WlO}.
26
1 Prologue
27
G.P. Thompson
(1)
C.T.R. Wilson
(2)
28
1 Prologue
world-conditions precisely in this way. A tensor express simultaneously the
whole group of measure-numbers associated with any world-condition; and
machinery is provided for keeping the various codes distinct.
And, just as in arithmetic we can deal freely with a billion objects without
trying to visualise the enormous collection; so the tensor calculus enables us
to deal with the world-condition in the totality of its apsects without attempting to picture it. {El, p. 3} (emphasis added)
29
30
these communities had died, and Leo would have had to move to states
such as Iowa or South Dakota to find ones that were still thriving, cf.
{AI}. But Leo's commitment ot Tolstoyism was not that strong: he
trampled on and wound up working in a department store in Kansas
City. According to Wiener, Leo did stumble across a Fourierist colony.
"It had gone to seed, and all the efficient poeple had left it, while all the
rogues and footless, incompetent idealists remained" [53h, p. 19].
Seeing a sign "Gaelic lessons given" on a church wall in the city,
Leo enlisted in the class and soon became its instructor, earning the
nickname "the Russian Irishman". A bit tired of his jack-of-all-trades
existence, Leo applied for a teaching position in the Kansas City school
system, and proved to be a successful and popular teacher. Later, he
moved to the modern languages department of the University of Missouri at Columbia.
To backtrack a little, the poet Robert Browning was the rage of
the ladies of Kansas City, and the erudite Leo was invited to speak at
their "Browning Clubs". There he met a Miss Bertha Kahn, daughter of
a department store owner of German-Jewish descent. In 1893 in Columbia, Leo, then a professor, and Bertha were married. Browning's verse
play On a Balcony has three characters, Norbert, Constance, and the
Queen. So when a son was born to Leo and Bertha on November 26,
1894, they christened him "Norbert", and when in 1898 a daughter was
born they called her "Constance". Who, if anyone was to be christened
the Queen, we shall never know, for a daughter born in 1902 was named
Bertha and a son born in 1906 was named Fritz. Unhappily, a son born
in 1900 and another in 1911 died in infancy.
While Bertha's father Henry Kahn, was a Jew, her mother came
from the Ellinger family, rooted in Missouri for at least two generations,
that was only partially Jewish. As Wiener tells us, "The family was
hovering in a state of unstable equilibrium between its Jewish background and an absorption into the general community" [53h, p. 23].
As for his mother Bertha, Wiener wrote:
She had been brought up with the indulgence often extended to the belle of
the family. I remember one photograph of her made when I was about four
years old. She looked extremely handsome in the short sealskin jacket of that
period. I had great pride in that picture and in her beauty. She was a small
woman, healthy, vigorous, and vivacious, as she has indeed remained to this
present day. She still carries herself like a woman in her prime.
In the family of divided roots and Southern gentility into which she
was born, etiquette played a perhaps disproportionately large part, and
trespassed on much of the ground which might be claimed by principle. It is
small wonder, then, that my mother had, and conceived that she had, a very
hard task in reducing my brilliant and absent-minded father, with his enthusiasms and his hot temper, to an acceptable measure of social conformity.
[53h, p. 27]
31
Bertha and
Leo Wiener,
parents of
Norbert Wiener
32
Wiener has left us with a vivid description of his father's personality, ideals and his disappointments from failure to attain them:
.. a small, vigorous man, of emotions both deep and quick, sudden in his
movements and his gestures, ready to approve and to condemn, a scholar
rather by nature than by any specific training. In him were joined the best
traditions of German thought, Jewish intellect, and American spirit. He was
given to overriding the wills of those about him by the sheer intensity of his
emotion rather than by any particular desire to master other people. [56g,
p.18]
.. he was essentially a German liberal of type well known in the middle and
later nineteenth century, fully in sympathy with the German intellectual
tradition as it had come down from the time when Goethe had been a truer
symbol of German aspirations than the Emperor Wilhelm II was ever to
become. Separated as my father was from Germany, largely self-educated,
and outside the orthodox German academic tradition, he still hoped for
many years that by sheer intellectual strength and integrity he could win from
Germany the recognition accorded to a great German scholar.
33
34
Wiener was fundamentally molded by the democracy of the New England of his childhood days. Later in life he came to value his sojourns
away from New England because they let him see "the great lines of its
spiritual character" from a distance [53h, p. 48].
For the rest, Wiener's childhood was shaped by the milieu
around his father's home. His Cambridge, Massachusetts, neighbors in
1901 were Maxime Bocher, the mathematician, and Otto Folin, the
physiological chemist, whose children became his playmates. The
Wieners had a friend in the great physiologist Walter Cannon of the
Harvard Medical School, and Norbert felt free to address his childish
questions to him.! His early biological curiosities were encouraged by
his father's mycological hobbies. He also met his father's interesting
friends: Prince Kropotkin, geographer and anarchist, whom he saw in
Cambridge and in London, and the British Zionist, Israel Zangwill, in
London, with whom his father often had arguments.
At six Norbert could read freely, and by eight he had overstrained his eyes and was threatened with myopia. Apart from the
Treasure Island type of fiction, he devoured books on science, especially
natural history. He kept reading, although as he recalls "much of it was
over my head" [53h, p. 65]. He "longed to be a naturalist as other boys
longed to be policemen or locomotive engineers". But shades of the
mature Wiener were already present in the child:
Even in zoology and botany it was the diagrams of complicated structures
and problems of growth and education which excited my interest as fully as
the tales of adventure and discovery. [53h, p. 64]
35
Peabody School in 1901, Leo realized that what the boy needed was not
the cribbing of multiplication tables and the like, but something interesting and challenging. So he removed Norbert from school and placed him
under his own tutelage with courses in algebra and classical languages.
This continued for about two years, after which Norbert reentered
school at age eight.
A critical idealist understands that effective action must accord
with the laws of nature, but often his earnestness gets the better of his
judgment and his temper. This was the case with Leo in his educational
dealings with Norbert. Norbert recalls how his fumbling in algebra or
grammar turned "a gentle and loving father" into "an avenger of
blood", and describes how his father's harsh "What?"s and "Do it
again"s terrified him and worsened his responses, only to enrage his
father even further and to make him hurl insults in German. The lessons
often ended in a family scene with a crying Norbert running into his
mother's arms. Pages 67~68 of [53h] bear eloquent testimony to the
destruction of the learning process caused by a teacher who equates
discipline with a mixture of harshness and irony, and resorts to the
public exposure of a child's stupidity, and to his harassment in the
presence of others.
Unlike the fathers of the prodigies John Stuart Mill and Samuel
Butler, Leo did not invoke any moral or religious sanction for his harsh
instructorship. Rather, he justified it on (somewhat questionable) educational grounds. In a 1911 article on "New ideas in child training", he
wrote:
It is nonsense to say, as some people do, that Norbert and Constance and
Bertha are unusually gifted children. They are nothing of the sort. If they
know more than other children of their age, it is because they have been
trained differently. [53h, p. 158]
What kept Norbert from falling into total despondency during this
difficult period, 1901~1903, was his image of his father as not just a
"taskmaster", but also a hero who was trying to attain the impossible
ideal for his son.
But Norbert never completely recovered from the strains of this
period. The differences he noticed between himself and other children,
stemming both from his precociousness and [rom his poor eyesight and
poor muscular coordination, and the wounds suffered by his ego during
conflicts with his father, turned him into an ego-defender who
sometimes treated the external world as a medium for ego-aggrandizement. At such moments, he lost his philosophical touch. His reaction,
at age 18, to his father's statement (quoted above) is an instance. All he
could see in it was the implication: "my failures were my own, but my
36
Norbert Wiener
at age seven
Fortunately the ban on reading was lifted after six months, and Wiener
was able to read freely until 1946, and then after an eye operation till the
end of his life. But he always preferred working on a blackboard to
working on a sheet of paper.
37
38
with "half an ear" while continuing his own work. "But half an ear was
fully adequate to catch any mistakes" that Norbert made and to provoke harsh reprobation [53h, p. 94].
Norbert graduated from the Ayer High School in the summer of
1906 at the age of eleven. During the three years he spent in its environs
he had made many friends and had many a happy moment. Years later
he expressed the meaning of this experience in his life as follows: "1 had
a chance to see the democracy of my country at its best in the form in
which it is embodied in the small New England town" [53h, p. 101].
Describing his revisit about forty years later to these same environs, by
then partially industrialized, he was to write:
I have the impression that my friends in this small industrial town represent
a sort of stability without snobbishness which is universal rather than provincial, and that the structure of their society compares well with the best that
a similar place in Europe would have to offer. When I go back among them
it is expected of me, and rightly expected of me, that I revert in some measure
to my status as a boy among the elders of the family. And I do so gratefully,
with a sense of roots and security which is beyond price for me. [53h, p. 101]
Wiener's love for the New England farm folk persisted until the end. He
expresses "eternal thanks" to his father for the property he acquired in
South Tamworth, New Hampshire, in 1909, [53h, p. 140], where Mrs.
Margaret Wiener, his widow, still resides. Of the New Hampshire farmer
he wrote: " ... whether you can love him or not, and very often you can,
you can and must respect him because he respects himself" [53h, p. 142].
39
The next most influential philosophical figure for Wiener was William
James, whose Lowell lectures on pragmatism he attended. But in spirit
Wiener was in many ways closer to the then unknown Charles Sanders
Peirce than he was to William James. Besides philosophy, Wiener's
extracurricular activities also included the study of biology.
At this stage Wiener began to display an innate ability in practical engineering design. He and his friends designed an electromagnetic
coherer for radio waves and an electrostatic transformer. He himself was
clumsy in the handling of apparatus, but felt at home in a laboratory
amid understanding colleagues who could deal manually with his ideas.
He also participated in biological experiments, but here guilt feelings
40
41
Wiener's first days at Cornell University were marred by another emotional crisis. He heard Professor Thilly of Cornell remark to his father
about the story that Rabbi Moses Maimonides was one of the distant
ancestors of the Wieners. This was the first time that the 15-year-old
Wiener realized that he was a Jew! This fact had been withheld from him
by his parents. The sudden realization that his mother had often spoken
derogatorily of Jews, that her maiden name, Kahn, was a variant of the
Jewish Cohen, and that she had lied when she denied his cousin's earlier
statement that they were Jewish were shattering:
The wounds inflicted by the truth are likely to be clean cuts which heal easily,
but the bludgeoned wounds of a lie draw and fester. [53h, p. 147)
42
"The black year of my life" was Wiener's epithet for the year 1911. With
much justice Wiener wrote:
Who was I, simply because I was the son of my mother and father, to take
advantage of a license to pass myself off as a Gentile, which was not granted
to other people whom I knew .... My protection may have been well intended, but it was a protection that I could not accept if I were to keep my
integrity.
If the maintenance of my identity as a Jew had not been forced on
me as an act of integrity, and if the fact that I was of Jewish origin had been
known to me, but surrounded by no family-imposed aura of emotional
conflict, I could and would have accepted it as a normal fact of my existence,
of no exceptional importance either to myself or to anyone else. [53h, p. 147]
Not only was this issue thrust into undue importance, but a crisis
involving intense suffering was engendered. Wiener sadly recounts his
immediate response to the crisis:
... I alternated between a period of cowardly self-abasement and a phase of
cowardly assertion, in which I was even more anti-Semitic than my mother.
[53h, pp. 148-149]
It was after considerable groping that Wiener learned to cope with this
issue. The agnosticism he had imbibed from his parents ruled out a
return to Orthodox Judaism, and his honesty forbade the complete
renunciation of his Jewishness. At last he saw the light, that anit-Semitism was just one manifestation of the much larger problems of human
prejudice and human exploitation. The consequence was a shift towards
enlightened liberalism [53h, p. 155].
During the 19th and early 20th centuries, many a Jewish family
accepted Christianity either from conviction or from expediency. The
Mendelsohn and Marx families in Germany and the von Neumann
family in Hungary are known examples. A remarkable case was that of
the French Jewish philosopher Henri Bergson, who wanted to convert
to Roman Catholicism in the late 1930s, but hesitated to undergo baptism
in view of the plight of European Jews under fascism. When Vichy
France promulgated anti-semitic measures in 1940, the enfeebled man,
close to death, stood in line to register as a Jew at his own request. Shortly
thereafter, as per his wishes, a Catholic priest prayed at his funeral. (At that
time the Church, in its supreme folly, had his books on its Index.)
The Wiener episode was markedly different. Here the parents,
while inculcating in their son a reverence for integrity, practiced dis-
43
44
Norbert Wiener in
1912 just before he
received his doctorate
from Harvard
Wiener received a Ph.D. from Harvard in June 1913. This happy event
was accompanied by the good news that he had been awarded an
overseas traveling fellowship by Harvard University, the John Thornton
Kirkland Fellowship.
As a result oftraining from his father, Wiener had a Ph.D. at the
age of 18, about six years below the average age. These valuable years
were at his disposal. This was a compensation for the suffering that
Wiener had to endure. The aggravation of the boy's difficulties by the
confusion of discipline with harshness and intimidation, and by misinformation about his Jewish origin, were tragic events. But Leo Wiener's
fundamental objective of keeping the boy on his intellectual toes, of
putting him in contact with good minds, of giving him the run of
libraries and laboratories, and a chance to love the countryside is
something for which we must all be thankful.
45
A Wiener's emancipation
Wiener decided to utilize his Harvard traveling fellowship to study
mathematical philosophy at Cambridge University under Bertrand Russell, on whose work his thesis had been based. Russell was Lecturer in
the Principles of Mathematics, and had just completed the monumental
and epoch-making Principia Mathematica { WII } together with his
former teacher, the philosopher Alfred North Whitehead. Wiener well
describes this phase of his life as "emancipation": he was to be in a
superb environment amid great minds, with the Atlantic Ocean acting
as a potential barrier to paternal interference. But it was only with his
father's intervention that this "emancipation" was procured. l Russell's
autobiography {R8, Vol. I, pp. 345, 346} contains a letter he received
from Leo Wiener. It is reproduced here in full, since it gives us a glimpse
of Leo's interesting personality, and the tale it tells has a touch of
humor. The emphasis we have added corroborates Norbert's complaints
about his father's overbearing attitutes.
29 Sparks Street
Cambridge, Mass.
June 15, 1913
Hon. B.A.W. Russell
Trinity College
Cambridge, Eng.
Esteemed Colleague:
My son, Norbert Wiener, will this week receive his degree of
Ph.D. at Harvard University, his thesis being "A Comparative Study of
Wiener's last sentence in [53h, p. 179], with the words "I then wrote to
Russell", is thus somewhat misleading. Also misleading is the short account
he gives of the first meeting he and his father had with Russell [53h, p. 183].
It transpires from Russell's letters that neither of the Wieners impressed him
favorably.
46
47
5A Wiener's emancipation
It was while taking Hardy's course that Wiener wrote his maiden paper
[13a] on transfinite ordinals. Wiener's warm relationship with Hardy
lasted till the latter's death in 1947. He probably knew Hardy better than
any other American mathematician, and he was invited to write Hardy's
obituary in the Bulletin of the American Mathematical Society [49fV
But Wiener's forte during this early period was in the foundations of mathematics rather than in its superstructure, and Russell
rather than Hardy was his guide, friend and philosopher on all matters,
except personal morality. Wiener did not approve of Russell's libertinism, and his witty remark on the philosophical libertine is worth recallmg:
There is a great deal in common between the libertine who feels the
philosophical compulsion to grin and be polite while another libertine is
making away with the affections of his wife and the Spartan boy who
concealed the stolen fox under his cloak and had to keep a straight face when
the fox was biting him. [53h, p. 192]
48
49
most sets are not members of themselves, some are. The set C of all cows
is obviously not a cow. So C is not a member of C; briefly, C $ C.
Likewise, for the set F of all finite sets, we see that each of the sets
{l, 2},
{I },
{I, 2, 3},
etc.
is finite and therefore a member of F. But there are infinitely many such
members, and so F itself is not finite; thus F $ F. On the other hand, the
set I of all infinite sets is infinite. For each of the sets
{l, 2, 3, ... },
{2, 3, 4, ... },
{3, 4, 5, ... },
etc.
if & only if
X $ X.
(1)
Given a set A, the formula (1) enables us to give a direct answer to the
question: Is A a member of R? It is, if A $ A; it is not, if A E A. Thus
for the sets C, F and I considered above, we find that C E R, FER but
1$ R. If, however, we ask the pertinent question 'Is R a member of R?',
we get into trouble, for the substitution of 'R' for 'X' in (1) yields the
contradiction
if & only if
R $ R.
(2)
50
51
Bertrand Russell
1872-1970
Photograph taken in
1916
52
53
2.
3.
54
55
and all were aware of the towering presence of Albert Einstein and of
his practice of the great tradition of blending deep physical intuition
with abstruse mathematics to secure very powerful hypotheses at the
empirical level.
This atmosphere affected Wiener's studies in a concrete way.
Russell urged him to approach mathematical philosophy from the
broadest standpoint, to concentrate not just on the foundations but also
to look at the frontiers of mathematics and theoretical physics. This
advice not only brought Wiener into contact with G. H. Hardy, as we
noted earlier, but it also exposed Wiener to Bohr's atomic theory, the
work of J. W. Gibbs on statistical mechanics and the Einstein-Smoluchowski papers on the Brownian motion, which were to prove of lasting
value, cf. 7D below. One wonders whether Wiener would have received
the wise counsel he obtained from Russell had he been a student in our
present fragmented and watered-down universities.
The work of the logical school in Cambridge (c. 1914) was
constricted by the absence of metamathematics. The only contribution
to axiomatic set theory that came from Wiener's contacts with Russell
was his theoretical definition of the ordered pair as a set [14a]. This
simplification, by completing a line of thought of Charles S. Peirce,
marked a new stage in the history of logic, in which relations and
functions had no longer to be treated as primitive terms. In Quine's
words:
In the reduction of logic in turn to the three present primitives,5 one
essential step was Russell's discovery of how to define complex terms in
context; a second was Sheffer's reduction of conjunction, alternation, and
denial to joint denial; and a third was Wiener's definition of the ordered pair.
{QI, p. 126}
Wiener's definition also took away some of the complexity of the type
theory in the Principia Mathematica, cf. {Ql, p. 163}.
This was no mean achievement for a lad of 19. But for reasons
unknown, Russell did not appreciate Peirce's simplification of relational
logic and still less that of Wiener. Russell has described Peirce as "one
of the most original minds of the late nineteenth century, and certainly
the greatest American thinker ever", cf. {R9, p.276}. But the folder
[MC, 468] contains a 1914 handwritten exchange in which Russell
comments, "I do not think a relation ought to be regarded as a set of
ordered couples", to which Wiener responds, "It seems to me that what
is possible in mathematics is legitimate". Wiener also wrote a paper [15b]
critical of the then prelavent view that a closed system of mathematical
logic is possible. This accorded with the intuitive feelings he had had
5
'E'
56
since childhood, and the subsequent work of G6del was to bear him out.
But in this he could merely surmise and not deliver.
Wiener also maintained contact with the other two philosophers,
besides Russell, belonging to the "Mad Tea Party of Trinity" , viz. J. M. E.
McTaggart, the Hegelian, and G. E. Moore, the analytic moral
philosopher. (These were, respectively, the Mad Hatter, the Dormouse
and the March Hare.) Wiener enjoyed the intellectual life of Cambridge
University. He has written glowingly of the interesting people he met at
Russell's parties, such as the young C.K. Ogden and I. A. Richards, just
then embarking on their The Meaning of Meaning. He had also made
friends with some of the senior students who were prospective dons,
such as F. C. Bartlett, later to become Professor of Psychology at
Cambridge. "Half-starved" though he was with the "raw carrots and
inedible Brussels sprouts" he got from his landlady, and with his own
"occasional penny bars of chocolate", he wrote "I was both happier and
more of a man than I had been before" [53h, p. 197].
Early in 1914, after spending Christmas with his family in Munich, Wiener wrote his paper [14b] wherein he adapted some ideas of
Whitehead to develop a relation of "complete succession" (which is
irreflexive and transitive but not total). This relation is well-adapted to
a theory of time and space in which entities such as instants and points
are regarded as constructs derived from psychologically less remote
entities such as temporal events having duration and spatial extension.
The construction of idealized scientific concepts from the more familiar
ones suggested by perception, by use of relation theory, is Whitehead's
principle of extensive abstraction. Whitehead had by then left Cambridge
for London, and Wiener went there to make his acquaintance. As
Russell was to be away during the May term of 1914, Wiener followed
his advice to study at Gottingen. Before his departure, however, he
completed an essay [14c] entitled "The highest good", which won him a
Bowdoin Prize from Harvard, and a paper [14d] on relativism.
Several ideas that were to 100m in Wiener's cybernetics (1948)
appear in [14c, d] in nascent form. In the more fundamental of these
papers [14d] Wiener denies first the position taken by F. H. Bradley and
G.E. Moore that self-sufficient knowledge and self-sufficient experience
are possible, and second he questions the possibilities of a priori knowledge or purely derived knowledge.
Taking what today we would call a metamathematical standpoint, Wiener emphasises the "grammar" of geometry, i.e. the rules
governing the formation of geometric sentences and the assignment of
meanings to the geometric terms 'point', 'line', etc. (For instance, 'point'
might mean a tiny dot, 'line' a stretched string.) This grammar has to be
formulated in a language MI. The grammar of M 1 will require for its
57
On Bergson's belief that these "fundamental dichotomies" are irreducible and sharp, Wiener wrote:
Now to suppose the existence of absolutely sharp distinctions runs directly
counter to the spirit of relativism and, I believe, of Bergsonianism itself. (14d,
p.570]
It was precisely the "gulfs" listed in the first quotation that cybernetics
bridged about 20 years later. Furthermore this bridging could be undertaken only after cyberneticists realized that the anisotropy of time
(Eddington's arrow), demanded by the laws of thermodynamics, had
more to it than what the physicists attributed: it also made room for
indeterminism, for teleological phenomena, growth, learning, and free-
58
dom-the very things that Bergson cherished but felt were missing.
Chapter 1 of Wiener's Cybernetics [61c] with the title "Newtonian and
Bergsonian Time" stresses this, while affirming that the absence of strict
determinism in the universe does not destroy its cosmic character.
Relativism also allows for the emphatic readings of absolutism,
pragmatism and Bergsonianism. Thus Wiener concludes that "all
philosophies are nascent relativisms" (p. 574). Wiener's remarks of what
we owe to our heritage, his analogy between the history of theories and
the history of tools (p. 576), his affirmation of the self-critical nature of
the scientific enterprise (p. 577), and his affirmation of the secondary
position of ignorance and noise-the physicist does not stop working
just because he can only arrive at approximate laws and can only have
instruments of imperfect accuracy (p. 575) - are in near-perfect harmony with the mature positions he was to take in later life.
In the essay [14c] Wiener rejects the view that gradation of
ethical behavior demands an ethical summum bonum, "the highest
good", for orderings can exist with no absolute maxima. " ... our ideals
grow with our attainments, ... the better a man becomes, the broader
are the vistas of righteousness that open out before him" [14c, p. 512].
Wiener then tries to separate sets of feelings that constitute conscience
from those that constitute prejudices. He answers the question as to why
individual consciences should concur, and bring forth a social conscience, by pointing to the biological impulse to survive, and to its
fulfillment by gregariousness, cooperation and concerted action. So
originates in us an instinctive urge to tolerate, if not respect, the consciences and prejudices of others. But individual consciences may antagonize one another, and the urge towards tolerance and towards a
common good is often overwhelmed by the warlike conscience. Force
thus becomes an arbiter between conflicting objective goods. Thus,
objective morality is the "end-product" of a two-fronted struggle, one
within the individual to reorder his ethical priorities, and the other between
groups of individuals. As this struggle is an ongoing one, objective morality
is far from being .fixed, and must evolve.
The thoughts expressed in [14c] reappear in more cybernetical
form in Wiener's theory oflearning machines that was to come 25 years
later, cf. 18E, F. All learning machines are dependant on feedback
with the environment, and to improve their future performance, they
must put to use the record of their previous performances-in short,
they must all learn from experience. So too for our species, "our ideals"
must "grow with our attainments" (p. 512). The view that "there is one
fixed immutable idea of moral conduct" (p. 512), or of any other kind
of conduct, ends the learning process at some point. Any such view is
obviously incompatible with Wiener's later thought, in which learning
59
In this letter the young Wiener clearly reveals the cleavage between the
tradition represented by Husserl, and we might add by mathematicians
prejudiced against the actual infinite such as Gauss, Kronecker and
Poincare, and that of the bolder school led by Cantor, Frege, Peano,
Whitehead and RusselU
6 Hyman Levy, later to become a professor of mathematics at London University, and a vocal member of the British left-wing.
7 "I protest against the use of infinite magnitudes as something completed,
which is never permissible in mathematics," Gauss, 1831.
"Later generations will regard Mengenlehre as a disease from which one has
recovered," Poincare, 1908, cf. {B5, pp. 556, 55S}.
60
Wiener's mind was far from idle during his Gottingen days, for
soon after his letter to Russell he completed a paper [I5a] on synthetic
logic which was to serve as the basis for his Docent Lectures at Harvard
the following year. He also tells us that here he had his first experience
of "the concentrated passionate work necessary for new research" [53h,
p. 211]. With his mind absorbed by matters foundational, the courses of
Hilbert and Landau did not register an immediate impact on Wiener.
But they had a delayed effect, and as his mathematical interests widened
he began to appreciate what he had learned from them. Wiener also
enjoyed the seminars and social life of the mathematical circle at Gottingen. He compared the situation there to that at Harvard by "a deep
draught of Muncher" as against "near beer". Among his companions at
the beer garden were the prospective mathematicians Otto Szasz and
Felix Bernstein.
Wiener returned to the United States in August 1914, for a
holiday. World War I had broken out, but his Harvard traveling
fellowship had been extended for the year 1914-1915, and so he
returned to wartime England. The atmosphere at Cambridge was
charged by the fervor of war. Russell's pacifism had made him an
unpopular figure, and the situation hindered undisturbed philosophical
investigation.
An interesting interlude for Wiener during this gloomy period
was a meeting with T. S. Eliot, great poet-to-be. Eliot had been a fellow
in philosophy and lndic philology at Harvard during the years
1908-1910 and 1911-1914, respectively, and knew Wiener. He came
under the influence of Bergson at the Sorbonne during 1910-1911, and
was at Merton College, Oxford in 1915 studying the idealist philosophy
of F.H. Bradley. He wrote to Wiener in October and December 1914
about a get-together. From the tone of these letters it is evident that he
knew Wiener quite well: "I can't imagine what on earth you are doing
with McTaggart unless you are reading Hegel or drinking whiskey"
[MC, 10]. They finally met in London, and with little money in their
pockets, had "a not too hilarious Christmas dinner together in one of
the larger Lyons restaurants" [53h, p.225]. They must have had a
considerable philosophical exchange, for there is a six-page letter to
Wiener from Eliot in London dated January 6, 1915, in which he
speaks of Wiener's papers [14a, b, c, d] and [15a], and comments at
length on the question of relativism in philosophy, broached by Wiener
in [14d). Eliot's interest is understandable in view of his own association with Bergson and Bradley. He also described his own embarkation on a critical thesis on Bradley's theory of judgment, Knowledge
and experience in the philosophy of F.H. Bradley. In this letter Eliot
wrote:
61
For me as for Santayana philosophy is chiefly literary criticism and conversation about life; and you have logic which seems to me of great value. [MC,
11]
(The entire letter is reprinted in the Coli. Works, IV, pp. 73-75.)
Although Eliot's description of philosophy as "chiefly literary
criticism" did not fit with Wiener's, the letter shows a commonality
between their flexible attitudes. Decades later there was to be a remarkable concordance between Wiener's views on the inevitability of the
collapse of homeostasis and Eliot's view on the inevitability of the
recurrence of sin: "Sin grows with doing good" {E6, p. 44}. One can only
surmise what influence, if any, Wiener had on Eliot's poetry.
E Columbia and its word-minded philosopher (1915)
As submarines were starting to surface in the Atlantic, Wiener's father
beckoned him to come home. Upon Russell's advice Wiener decided to
spend the rest of his traveling assignment at Columbia University in
New York, possibly under the philosopher John Dewey. This was
perhaps the only unfortunate advice that Wiener got from Russell.
Wiener arrived in New York in January 1915 and soon regretted
coming to Columbia University:
I found the skyscraper dormitories of Columbia depressing after Cambridge
and Gottingen. I also found the life of the place unsatisfactory in its lack of
coherence and unity. Almost the only bond between the professors, who
lived widely scattered in University Heights apartment houses or in suburban
bungalows, was an almost universal antagonism toward Nicholas Murray
Butler8 and everything that he stood for. [53h, pp. 221-222]
62
63
underway only in the early 1920s with the work of Alexander and
Veblen, both of whom Wiener was to meet a couple of years later at the
Aberdeen Proving Grounds of the U. S. Army. Thus Wiener was five
years ahead of the time in his ventures. These were not systematic,
however, and no pioneering contribution resulted. Unfortunately for us
Wiener lost his manuscript sometime before 1920.
Wiener, writing for the lay reader, has described combinatorial
topology (analysis situs) as:
... that strange branch of mathematics dealing with knots and other geometric shapes whose fundamental relations are not changed even by a
thorough kneading of space so long as nothing is cut and no two remote
points of space are joined. Topology includes the study of such things as the
familiar one-sided Mobius's sheet of paper, which you get when you take a
long, flat strip, rotate one end of it through half a revolution and glue the
ends. It makes an excellent parlor trick to ask a layman what will happen to
such a strip if you start cutting it down the middle until the ends of the cut
meet. If you try this, you will find that even after the cut is complete the strip
will remain in one piece but will now make a full revolution instead of half
a revolution as you proceed around it. [56g, pp. 26, 27]
64
Wiener's assignment at Harvard was temporary. While his accomplishments were remarkable for a 21-year old, they were not always rigorous,
as we just saw, nor were they substantial enough for a Philosophy
Department of the stature of Harvard to consider him for a long-term
appointment. Nor was Wiener a good teacher of undergraduate routine,
65
and this may have precluded the short-term renewal of his appointment.
Also, there is truth in Wiener's contention that "in those days work in
mathematical logic was equally unhelpful in getting jobs in either mathematics or philosophy" [53h, p. 270]. Wiener had to seek a job elsewhere,
and on his father's suggestion he tried to get one in a mathematics
department. With the aid of a teacher's agency Wiener got an instructorship in mathematics at the University of Maine in Orono, Maine, and
began his duties there in the fall of 1916.
Words such as 'dull routine', 'examination cribbing' and 'copying of homework' occur in Wiener's description of the academic environment at Maine. What made him 'feel alive' were contacts with the
statistician Raymond Pearl and his group, and getting off campus to
Bangor. He left the University of Maine at the end of the academic year.
Note: The Peano postulate system
This system for the set N of natural numbers 1,2,3, ... has two
undefined terms 1 and S, i.e. successor, (S(n) is to mean "the successor
of n", e. g. S(7) = 8). It has four postulates:
(1)
1 E N;
(2)
(3)
(4)
if n E N, then Sen) E N;
for m, n E N,
SCm) = Sen) if and only if m = n;
if E <;; N, 1 E E, and nEE implies Sen) E E, then S = N.
66
67
One has to read Wiener's letters during this period to realize that
he had become something of a "warmonger". A "Dear Dad" letter from
Orono, Maine dated March 1, 1917, about a month before the United
States' entry into the war, is especially pungent:
... I have the misfortune to have a hostess who is both suffragette and
pacifist-the yellow banner and the yellow streak. I wonder how long Wilson
will continue to turn our other cheek? I am afraid that the West is suffering
from hypertrophy of the pocket-book and atrophy of the conscience. It is a
sad thing for a nation when to epithmetikon assumes control over nous and
thumos, to use the language of the Republic.
Have you run across any good anathemata among your medieval
documents? I am getting a little short of cusswords for use against the
Germans and their American allies. [MC, IS]
The letter ends with the wish to be "back in the good old Harvard
Regiment:-as corporal of the 12th squad, company B". One gets a less
exhilarating picture of Wiener's attitude to the war from his 1953
account of these events in [53h, p. 243].
Having failed in all his efforts to join the armed forces, Wiener
searched for ajob related to the war effort, and in the fall of 1917 got
an apprenticeship in the General Electric Factory at Lynn, Massachusetts. He helped run some combustion tests, using a bit ofmathematics in thermodynamic calculations. He felt like an "honest workman".
"tired but happy". But his father, who knew the managing editor of the
Encyclopedia Americana, then in Albany, New York, secured a job for
him there as a staff editor, and Wiener abruptly terminated his services
at GE. In 1953 he confessed to his spinelessness in this act:
Though I felt morally obligated to stay on at the General Electric Company,
I was too dependent on my father to dare to contravene his orders, and so
I had to present my shame-faced resignation to the engineers who had given
me my chance in Lynn. [53h, p. 248]
68
and IV.) He even toyed with the idea of collecting some of them into a
book. Wiener has described his experience in this phase of his life as
follows:
With all the shortcomings and unpleasant sides of hack writing, it was a
wonderful training for me. I learned to write quickly, accurately, and with a
minimum of effort, on any subject of which I had a modicum of knowledge.
[53h, p. 251]
But he realized that despite the enjoyable and educative aspects of such
a position, it could not be the terminus of his intellectual activity. He
looked for a way out, preferably for one that led to the war effort. The
way opened when he received a letter from Professor (then Major)
Oswald Veblen, dated July 5, 1918, offering him the position of computor at the U. S. Army Proving Grounds in Aberdeen, Maryland, at a
salary of$1200 per year with 30% for overtime. Wiener readily accepted
after resigning his position at Albany, with high hopes of making
himself useful in the war effort.
At the Proving Grounds Wiener joined in the construction of
range tables for the new types of artillery and ammunition that were
being designed. The demands for speed and accuracy necessitated the
use of computing machines. As Wiener put it:
It was a period in which all the armies of the world were making the
Wiener felt that much significant mathematical activity during the postwar years emanated from those disciplined at Aberdeen: he cites Veblen,
Bliss and Alexander, among others. He speaks glowingly of his contacts
there with Hubert Bray and Philip Franklin (later to become his brother-in-Iaw and colleague) and others. He wrote: "Whatever we did, we
always talked mathematics", and
... I am sure that this opportunity to live for a protracted period with
mathematics and mathematicians greatly contributed to the devotion of all
of us to our science. [53h, p. 258]
But even this active involvement with war work was not enough
for Wiener, for in October 1918, he enlisted as a private in the U.S.
Army. Once in he began to regret it, for the war was nearing its end and
69
Norbert Wiener (far right) in army uniform at Aberdeen Proving Grounds, 1918
Even with a temperament not suited for a regimented life and a more than
average desire to sec what I was doing and to know what it meant, I had
found a few months of army life a haven at a time when I had been very tired
for years from making my own decisions. It has been said many times that
the motives of the soldier and the monk are curiously similar. [53h, p. 261)
Wiener's stint as an army private had its painful moments, but also its
funny side. During sentinel duty, armed with a rifle and bayonet, he
"found it hard not to drowse off and to keep sufficiently alert to
challenge the Officer of the Day" [53h, p. 260].
70
71
72
7C Postulate systems
73
C Postulate systems
It is worthwhile to hear Wiener's own semi-popular description of the
subject of postulation:
The geometry of the Greeks went back to certain initial assumptions, known
variously as axioms or postulates, which were conceived to be unbreakable
rules of logical and geometrical thought. Some of these were of a predominantly formal and logical character, such as the axiom that quantities equal
to the same quantity must be equal to each other. Another, with a more
purely spatial content, was that known as the parallel axiom, which asserts
that ifwc havc a plane containing a line t and a point P not on that line, then
through P and in that plane one and only only one line can be drawn that
will not intersect t. [56g, p. 51]
74
A the set of Asians, B the set of boys, and C the set of Chinese, then
A U B is the set comprising all Asians and all boys, A n B the set of all
Asian boys, and A' the set of all non-Asian people. Also since all Chinese
are Asians, C s: A. In the postulational treatment, we forget all about
sets, and instead lay down a few postulates such as (A U BY = A' n B'
(which are fulfilled by sets) from which all the other propositions can be
deduced.
In 1904 Huntington had given a set of nine such postulates for
a Boolean algebra A, governing U, 0, 1, " which were the only specific
primitive terms. He could then define the other operations such as n, s:,
etc. Wiener asked if operations other than union could satisfy the same
nine postulates. Towards this end, Wiener used Boole's result that every
binary operation Ell for A has the form
x Ell Y = (a
n x n y)
U (b
n x n y')
U (c
n x' n y)
U (d
n x' n y'),
Now Wiener had been strongly impressed by I. M. Sheffer's 1913 discovery (actually re-discovery of an 1880 unpublished result of C. S. Peirce)
that a single connective t (,neither-nor') suffices for truth-functional
sentential logic. For instance, the sentence 'p or q' can be rendered
'(p t q) t (p t q)', and similarly for 'and', 'not', 'only if', etc. In [20b], in
the same spirit, Wiener introduced a single binary operation (which we
shall write '*') subject to 7 postulates, and showed the equipollence of
this system having just one connective * to the usual one for a field IF.
Wiener's x*y is our familiar 1 - (x/y); e.g. 3*5 = 2/5,5*3 = -2/3. But
3*0 has no meaning, and thus * is not strictly a binary operation for IF.
(For details, see Note 2, p. 89.)
Wiener's spare time reading in 1919 included the works of the
distinguished French mathematician Maurice Frechet, and soon his
7C Postulate systems
75
76
Such steps or arrows are called vectors. Note that they can be of different
lengths. The length of a vector x, also called its norm, is denoted by Ixl.
Wiener's postulational work pertained to spaces (i.e. sets) Vofnormed
vectors.
To see the significance of the norm, we must go back to 1905,
when Frechet revealed the profundity behind the age-old idea of distance. He introduced metric spaces, i.e. spaces X for which it makes
sense to speak of a numerical distance d(x, y) between any points x, y in
X. Frechet showed that the theory oflimits, worked out during the 19th
century to rigorize the calculus of Newton and Leibniz, extends to
metric spaces. The distance (also called metric) of course serves as a
measure of proximity: the smaller the distance d(x, y), the "closer" are
the points x and y. Thus metric spaces are topological spaces, of an
attractive sort in which proximity is numerically measurable. Now the
Pythagoreans knew that many numbers, such as V2 and n are "irrational", i. e. are not equal to fractions min, where m, n are integers. Cantor's
way of incorporating these irrationals into the number system was to
postulate that any numerical sequence (x)' x 2 , x n, ... ) for which
Xm - Xn tends to zero, as m, n tend to infinity, has a limit t. Frechet
called a metric space X complete, when it fulfilled the anologous Cantori an postulate, i.e. when each sequence (x), x 2 , x n, ... ) in X, for
which d(xm' x n) tends to zero, has a limit t in X.
Now notice that in a system V of normed vectors, we can
introduce the distance dby the formula d(x, y) = Ix - y I. Thus normed
vector spaces are metric vector spaces.
What Wiener did in the papers [20e] and [22c] was to consider
affine normed vector spaces X: to each pair x, y in X corresponds a vector
xy in V, where V is a normed vector space of arbitrary dimension.) In
these papers he gave postulates for Vand then for X. Such vector spaces
were arousing a lot of curiosity at that time. Only in his paper [23g],
however, did Wiener impose the requirement that the metric defined by
his norm be complete. S. Banach, a Polish mathematician, later to gain
great stature, had, however, anticipated Wiener by a few months in his
own paper, which Wiener cited in the paper [23g]. The spaces are
therefore called Banach spaces (or "espaces du type B"). Wiener dealt
with Banach-space valued functions of a complex variable and extended
classical results, but he established no deep results of the sort that
Banach was proving. His explanation as to why he did not pursue the
Tn [22c, 7) Wiener also defined a local ("im kleinen") affine space, i.e. in
current terms, a locally pre-Banach ian topological manifold, but he did nothing with it beyond citing the sphere and the torus as examples. In [22b) he
also introduced "angles" in the space t 2' but did not pursue this idea.
77
So let us not start speaking of "espaces du type BW", but just note in
passing Professor R. L. Wilder's words:
The technical and conceptual competence exhibited in these papers leaves no
doubt ... that he (Wiener) would have been highly successful had he chosen
to continue. {Call. Works, I, p. 319}
78
Lebesgue integral. His 1913 course with G.H. Hardy, his third great
mentor, at last became relevant to his research.
Wiener has described the Lebesgue integral and its uses in lay
terms as follows:
... It is easy enough to measure the length of an interval along a line or the
area inside a circle or other smooth, closed curve. Yet when one tries to
measure sets of points which are scattered over an infinity of segments or
curve-bound areas, or sets of points so irregularly distributed that even this
complicated description is not adequate for them, the very simplest notions
of area and volume demand high-grade thinking for their definition. The
Lebesgue integral is a tool for measuring such complex phenomena.
The theory of the Lebesgue integral leads the student from the
measure of intervals to the measure of more complex phenomena obtained
by combining sequences of intervals, and then to sets which can be approached by such sequences, while the sets of points excluded from them can
be approached in a similar manner. ... It enabled Lebesgue to extend the
notion of length or measure from the single interval to the extreme significant
limits at which measure is possible. [56g, pp. 22~23l (emphasis added)
Wiener's intent was to extend Lebesgue's ideas so that they could apply
to "intervals" each "point" of which is an erratic curve, such as the orbit
of a flying bee. This was too difficult to do directly.
Fortunately, Wiener was aware of the 1919 paper of the British
mathematician P.l. Daniell {Dl}. Daniell had observed that in essence
Lebesgue began with an integral S defined over a class 2 of conveniently
chosen "elementary functions" (e.g. the step-functions or the continuous functions), and with these ingredients then built a much larger
and mathematically more beautiful class 1 of functions, and extended
the integral, so that the extended integral over 1 had richer properties. Daniell gave postulates that exemplified the Lebesgue procedure in
the abstract. He considered a system (2, M), where 2 is a vector space
of real-valued functionsf, g, etc. over a set X, i.e. 2 is such that.
f, g,
2 & a, b
IR
implies
af
bg
2,
79
80
A. Einstein
(c. 1910)
1879-1955
Some will push in one direction and some in another, and the balance of
pushes is likely to be tolerably even. Nevertheless, notwithstanding these
balanced pushes, the fact remains that they are pushes by individual people
and that their balance will be only approximate. Thus, in the course of time,
the ball will wander about the field like the drunken man ... and we shall
have a certain irregular motion in which what happens in the future will have
very little to do with what has happened in the past.
Now consider the molecules of a fluid, whether gas or liquid. These molecules
will not be at rest but will have a random irregular motion like that of the
people in the crowd. This motion will become more active as the temperature
goes up. Let us suppose that we have introduced into this fluid a small sphere
which can be pushed about by the molecules in much the way that the
pushball is agitated by the crowd. If this sphere is extremely small we cannot
see it, and if it is extremely large and suspended in a fluid, the collisions of
the particles of the fluid with the sphere will average out sufficiently well so
that no motion is observable. There is an intermediate range in which the
sphere is large enough to be visible and small enough to appear under the
microscope in a constant irregular motion. This agitation, which indicates
the irregular movement of the molecules is known as the Brownian motion.
81
J.B. Perrin
I 85()""1 942
Courtesy of
Dr. Francis Perrin
82
1905 paper, stating that "if the movement discussed here can actually be
observed ... an exact determination of atomic dimensions is then possible" {E2, pp. 1, 2}. From the atomic hypothesis and the principle of
equipartition of energy, Einstein deduced that at temperature T the
mean-square displacement, at> during a time-interval of length t, of a
colloidal particle of radius a (such as Brown's pollen grain) is given by
d/
~. t,
3naflN
(1)
where R is the constant of the gas equation (pv = RT), fl is the viscosity
of the liquid and N is Avogadro's number. 3 From this formula, N is
expressible in terms of experimentally observable quantities and known
constants. The relevant experiments were made by several physicists,
among them J. Perrin. These yielded a value of N which matched its
value obtained earlier by the entirely different method of diffusion of
ions in a gas due to an electric field. This was a brilliant confirmation not
only of the soundness of Einstein's research, but of the molecular-statistical theory itself.
This illustrious history may have encouraged Wiener to plunge
into the study of the Brownian movement. But with his philosophical
training, concern with sensation-intensities and their measurement (witness his papers [l5a, 21aD he must also have sensed the ubiquitous
character of the Brownian movement. Not just the pollen grains get
kicked around by the ceaseless molecular motion, so too do the electrons, whose flow in electrical apparatus control the transmission of
messages, and the colloids, whose flow in biological organisms control
life and mind. This realization of a randomness in the very texture of
Nature profoundly affected Wiener's life work, as we shall see in this
book.
Einstein had assumed that there is a positive number T such that
a time-interval of length T is, in his words
.... very small compared to the observed interval of time, but
nevertheless, ... such ... that the movement executed by a particle in consecutive intervals of time Tare ... mutually independent phenomena. {E2,
p. 13}
From this premise he derived the result that the displacements in disjoint
intervals are normally distributed, that "the mean (square) displacement
83
is ... proportional to the square root of the time" {E2, p. 17}, being
given by equation (l).
Wiener's concern, unlike Einstein's, was with "the mathematical
properties of the curve followed by a single particle" [56g, p. 38]. He
therefore made the idealization that Einstein's conditions prevail for all
positive lengths T. To this idealized Brownian motion, "an excellent
surrogate for the cruder properties of the true Brownian motion" [56g,
p. 39], Wiener was able to apply the theorem proved in [20f]. Tn his first
and remarkable paper [23d] Wiener followed a different approach, but
the nexus with [20f] is clear from 3, 4 of the paper [24d]. Here Wiener
defined the sequence (n", Wn)~l with X = [0, 1], so that it not only fulfills
the premisses of the [20f] theorem, but the extensions (2, M) of the
resulting (fe, M) also have the following additional properties. Write
{
[R
~(f)
84
terize it as the process for which x(t, .) is normally distributed with zero
mean and such that
[0, I],
(4)
where s /\ t means the minimum of sand t.
Wiener showed that for almost all a in [0, I], the trajectories
x( " a) are continuous everywhere but differentiable nowhere, thus
confirming what Perrin's microscopic observations had suggested.
Although the functions x( . , a) are of "extreme sinusoidity", and nonrectifiable (i. e. even tiny pieces have infinite length), Wiener was able to
define for any f in L 2 [a, b], a "Stieltjes" type integral
S>(s, a)x(t, a)da = s /\ t,
g( .) = I:J(t)dx(t, .)
s, t
on [0, I],
(5)
occur in the later works [33a, 34a, 34d] done in collaboration with Paley
and Zygmund.
In the theory of stochastic processes the idealized Brownian
motion occupies a central position, as the efforts of several mathematicians over the last fifty years have shown. Wiener may have surmised
this, but he did not realize that his Brownian motion also penetrates
deeply into the non-stochastic parts of mathematical analysis. A very
interesting example of this is afforded by the initial value problem of the
(one-dimensional) heat equation with potential term V:
u( ,0) = rp(')
on (~oo,oo),
au
at
I au
'2 af ~
2
V(x,t)u,
t:2: 0.
.) of this is given by
S e -S;V{x + f(r),t-t}dr:. rp{x
+ J(t)}w(df),
qQ,oo)
85
Despite their depth and wide scope, Wiener's papers on Brownian motion did not create any stir in the mathematical world. The
leaders of American mathematics at that time were G.D. Birkhoff of
Harvard, a Poincare analyst, and Oswald Veblen of Princeton, a topologist. They were interested in physics, but primarily in the then exciting
areas of general relativity and quantum theory; their horizons did not
86
87
88
X(O, IX)
x,
YeO, IX)
= y,
IX E [0, 1],
and let rex, y, IX) be the first time that this planar Brownian curve
(X( . , IX), Y( . , IX)) leaves R. Then the generalized solution of the Dirichlet problem
Llu
= 0 on R,
= rp on
89
7 Notes
In [17a] Wiener showed that the general binary operation $ will satisfy
Huntington's very axioms, if and only if in Boole's expression d = 0, c
= b and a = 1; with such a choice for a, b, c, d, the expression for $
reduces to
x
For this
Y = (x
$,
n y)
U {b
n (x
U y)}.
n y)
U (E
n (x u y)},
but the new complementation is still -. Wiener then showed that the
one-one correspondence
x
--+ Xl
= (a
n x)
U (a
n x)
'y
(x
n y)
a,
U (a
n (x
U y)},
so that 1',0' are now a. In this way Wiener was able to sum up the
invariant theory of Boolean algebras.
Note 2: Algebraic field [20b] (cf. 7C).
With his postulates on the sole binary operation*, Wiener showed that
o and 1 are uniquely defined in IF by x*x and O*y, y of. O. He defined
multiplication for (non-zero) x, y, by
X
y = (1*{[{(1*y)*I}*x]*I})*l.
A
E
+
+
Ex
Fx
+ Cy + Dxy
+ Gy + Hxy
----------~----~
with rational number coefficients, that do the same job. In [20c,d, 21 b],
Wiener defined two connectives * and # to be equivalent, if each could
be had by an iterated application of the other, and he studied conditions
for equivalence, including in this the complex field. This last field,
denoted by C, comprises all numbers of the form x + iy, where x, yare
ordinary real numbers, and i = V( -1).
90
7r n -simple},
F(t)wn(A)}/{ L wn(A)}.
tEL1E1t
LIEn:"
PlJ(:Jl').
91
7 Notes
q+~}
92
Wiener's competitiveness, as revealed by this incident, had something to do with his slow promotion in the Mathematics Department at
MIT. Between 1919 and 1924 Wiener did some of the best mathematical
work of this century. During this period he received considerable human
help from his MIT colleagues, but all he got officially from MIT, in
1924, was an assistant professorship without tenure. One might explain
this circumstance by saying that it happened in an era of generally slow
promotions, before the advent of President Karl Compton with his
enlightened policy on the status of science at MIT, and that Wiener's
pioneering work on Brownian movement was about ten years ahead of
its time and few understood it.
To say this, however, is to say little, for in reality tension over
promotion looms large because the scientists have to share a tiny fragment of the national income, consequent to its large-scale diversion in
support of inefficient bureaucracies, substandard and questionable manufacture, and pseudointellectual culture. Unfortunately, Wiener did not
analyze the problem objectively, and see it as the mal-allocation of
93
Thus the intellectual rockets that uplifted Wiener's research and uplifted
English poetry were products of injustice. The authors of Edwardian,
Elizabethan and Tudor gossip reaped rich rewards as do our pornography queens. "Values are for the birds," said the devil with a chuckle.
94
Marriage
95
Marguerite Engemann
in 1926, the year of
her marriage to
Norbert Wiener
O. Veblen, but nothing came of it. Finally, in 1929 MIT promoted him
to an associate professorship. He might have remained in this intermediate category for another eight to ten years, were it not for the fact
that his forthcoming researches, although inspired by the needs of
engineering, put him in command of a well-established purely mathematical area, head and shoulders above those inside the establishment.
This led to his promotion to a full professorship in 1932.
In this last research Wiener's scholarly pen reigned supreme:
nothing was forced and egoism was gone. We must now turn to this
phase of Wiener's life, in which he again became a true practitioner in
the Pythagorean-Platonic tradition.
96
A Harmonic analysis
Harmonic analysis, for all its modern ramifications, has a history going back
to Pythagoras and his interest in music and the vibrations of strings in the
lyre. [56g, p. 105]
97
9A Harmonic analysis
ao
2
k~l
{akcos(knt)
ak
= -1
T_T
bksin(kTnt)}.
ak,
(1)
b k = -1
SJ(t) .
T
knt dt.
Sill -
-T
(2)
i sin },t,
(3)
This last expression e;Al is justified by the connection that Euler found in the
18th century between the sinusoids and the transcendental real number e =
2.71828 ... with a non-terminating decimal expansion.
98
L Ckeik(n/Tlt
f(t)
&
Ck
k~ -00
-1
2T
JT
f(t)e-ik(n/T)tdt.
-T
(4)
f(t)
00
L j(k)eik(n/T)t,
j(k)
k~ -00
_1
2T
JT
f(t)e-ik(n/T)tdt.
-T
(5)
To put Wiener's words into symbols, it will suffice to give only the
complex number formulation. The equations (5) now give way to
f(t)
_1_
V(2n)
Joo
-00
j(A)eiAtdA, j(A)
_1_
V(2n)
Joo
f(t)e-iJ.tdt. (6)
-00
99
9A Harmonic analysis
(7)
the last being a Lebesgue integral not with respect to the length measure,
but with respect to F, which we may think of as the mass distribution
along a stick laid out on the A axis. By decomposition, such an integral
can be defined when F is the difference of two increasing functions and
even when it has an imaginary part of the same sort. F is called the
(indirect) Fourier-Stieltjes transform of F.
Finally, all the previous remarks extend to higher dimensional
spaces. Thus, for ordinary 3-dimensional space ~3, we take two points
A and x,), = (..1,1' ..1,2, ..1,3), X = (Xl' X 2, x 3) and define the character e). by
(8)
as(f)
100
This shows that the characters are absolutely intrinsic to the study of the
phenomena governed by time-invariant linear systems, and the problems of expressing functions as combinations of characters, and finding
the "Fourier coefficients" c, for a given/, in short, harmonic analysis and
synthesis, are exceedingly important.
B Heaviside's operational calculus [26c, 29c]
The first task that Wiener undertook at the behest of his engineering
colleagues was the rigorization of the 1893 operational calculus of Oliver
Heaviside. As Wiener wrote: "The brilliant work of Heaviside is purely
heuristic, devoid even of the pretense to mathematical rigor. Its operators apply to electric voltages and currents, which may be discontinuous
and certainly need not be analytic" [26c, p. 558]. The thought underlying
Wiener's rigorization is that "when applied to the function e nit , the
operator f(d/dt) is equivalent to multiplication by f(ni)" [26c, p. 560].
Given an arbitrary function/, he dissected it into a number of frequency
ranges, and applied to each range that expansion of f(d/dt) which
converged on this range.
By making these moves, Wiener was in effect groping towards
the theory of distributions that was to come twenty-five years later with
Laurent Schwartz. In 8 of his paper, which deals with the operational
solution of second order linear partial differential equations in two
variables, Wiener wrote
... there are cases where u must be regarded as a solution of our differential
equation in a general sense without possessing all the orders of derivatives
indicated in the equation, and indeed without being differentiable at all. It is
a matter of some interest, therefore, to render precise the manner in which
a non-differentiable function may satisfy in a generalized sense a differential
equation. [26c, p. 582]
In this he not only anticipated Professor Schwartz, but as the latter tells
us, in 1926 Wiener had seen farther than what all others were to see until
1946:
101
It est amusant de remarquer que c'est exactement eette idee qui m'a pousse
moi-meme it introduire les distributions! * Elle a tourmente de nombreux
matMmaticiens, comme Ie montrent ces quelques pages. Or Wiener donne
une tres bonne definition d'une solution generabsee; fen avais, dans mon
livre sur les Distributions, attribue les premieres definitions a Leray (1934),
Sobolev (1936), Friedrichs (1939), Bochner (1946), la definition la plus
generale :tant celte de Bochner; or la definition de Wiener est la meme que
celie de Bochner, et date done de ce memoire, c'est-a-dire de 1926, elle est
anterieure a toutes les autres. {Call. Works, II, p. 427}
f1
f2 .on ( -
00,
implies
t)
It f.oll.ows easily that the transfer .operat.or .of a time"invariant linear filter
with c.onv.oluti.on weighting W, i. e. the filter which yields f.or inputfthe
.output g:
g(t) = (W*j) (t) =
roo W(t-s)f(s)ds,
(W*f) (t)
O.on ( -
00,
(10)
LooW(t-s)f(s)ds,
real.
(11)
102
Thus Wiener found in the voltage curve of a busy telephone line the
same kind of local irregularity and overall persistence that he had
encountered in the Brownian motion, and he began to associate the
communication phase of electrical engineering with the statistical phase
of mechanics. For these phases, new and more difficult mathematical
methods were required. He set about to find them, spurred on by his
engineering colleagues.
D Generalized harmonic analysis
Wiener felt that signals of wide varieties should be harmonically analyzable, and that for this, the wider class of irregular and persisting curves
must be properly demarcated by use of new averaging operations. In this
research the earlier work of professional mathematicians on the rigorous
study of various classes of functions having convergent Fourier series
and integrals was of little use. Wiener therefore had to seek his ideas
from non-establishment "radicals" such as Lord Kelvin, Lord Rayleigh,
Sir Oliver Heaviside, Sir Arthur Schuster and Sir Geoffrey Taylor, who
had been interested in the harmonic analysis of allied random phenomena in acoustics, optics, and fluid mechanics. Another radical on whom
Wiener could have leaned was Albert Einstein, but his 1914 work in
harmonic analysis {E3} came to light only in 1985 (see Note 2, p. 112).
Building on the ideas of G. I. Taylor, Wiener introduced the class
S of measurable functions f on the real axis IR for which the covariance
function rp:
103
Lord Kelvin
1824-1907
Lord Rayleigh
1842-1919
104
rp(t)
lim - 1 JT f(s
2T - T
T->oo
+ t)f(s)ds
(12)
rp(t)
[00 eitAdF(A),
cf. 9A(7).
(l3)
F is called the spectral distribution off In the special case in which rp has
a finite Lebesgue integral, (13) gives way to
(14)
i. e. r is just the Fourier transform of rp. The first to use a function such
as F' was Schuster. At the turn of the nineteenth century, he used the
size of r().) to locate the "periodicities" "hidden" in the erratic function
f Schuster called r the periodogram off For the justification of (13)
and the historical antecedents of (12), (13) and (14), see Note 2, p. 112.
Since the function rp averages out the irregularities off, and is
rather smooth, it is to us not surprising that rp has a Fourier-Stieltjes
transform; indeed, Schuster himself had an inkling of this development.
But Wiener did not stop here. He went on to define a Fourier associate
s of the extremly irregular function f itself by the complicated formula:
SeA)
1
v(2n)
hm. --:;A->co
+-
-zt
V(2n)
eitA _1
f(t) - - d t .
-1
-it
(15)
This s is called the generalized Fourier transform off This function had
no historical antecedents, and was exclusively Wiener's creation. He got
both F and s in his memoir [30a], after much groping extending back to
[25c]. Now a well-known result in classical harmonic analysis is the
Bessel identity between a function f and its Fourier transform J, viz.
9E Tauberian theory
105
(16)
T~XJ
IT
II(t) 1 2 dt
lim - 1
h~O 4nh
-T
oo
I seA
-00
and he showed that its correctness hinged on that of the simpler identity
for non-negative g:
lim -1
T ~oo T
ITg(t)dt
0
lim -2
h~O nh
I get) -2-dt.
sin ht
2
oo
(18)
But Wiener had a hard time proving the latter, and here matters stood
in 1926.
E Tauberian theory
4 The invitation to lecture had come from interest in the quantum mechanical
aspect of another lecture he had delivered in Gottingen in the summer of
1925 (cf. Ch. 10).
106
G.H. Hardy
1877-1947
Photograph taken
around 1910
would reduce the integrals occurring in Schmidt's theorems to convolution integrals with which he was familiar from his electrical studies,
cf. 9B (10). Thus lettingf(u) = lfJ(e
g(u) = e-ulf!(e-"), we have
U
),
107
9E Tauberian theory
roo Wet -
s)f(s)ds,
real.
(19)
[00 Wet -
s) f.1 (ds).
(20)
All the classical Tauberian theorems, even the very deep, can be recovered from Wiener's two theorems: one has only to pick W, Wo,fand
f.1 intelligently and change variables. This applies also to (18) and
therefore to the Bessel-type identity (17) on which rests the appropriateness of Wiener's generalized Fourier transformation. Thus generalized harmonic analysis is put on a sound footing. But many other
theorems are also uncovered. Among these, perhaps the most interesting
is the one on the Lambert series, which bears on the analytic theory of
prime numbers. It led Wiener and his former student from Japan,
Professor S. Ikehara, to a simpler proof of the celebrated Prime Number
Theorem, to wit,
. n(n) . log n
11m
n
1,
n-'J.X!
where n(n) is the number of primes not exceeding n. For this proof, all
one had to know about the Riemann zeta function is that it has no zeros
108
on the line x
1 in the complex plane, and the traditional appeal to
contour integration was avoided.
Wiener's friend, J.D. Tamarkin of Brown University, induced
Wiener to gather his scattered results and present them systematially in
two memoirs. In 1929 Wiener wrote the first memoir [30a] on G.H.A.
In 1932 appeared his great memoir [32a] on Tauberian theorems. In the
final organization of both memoirs Wiener received considerable assistance from both Professor Tamarkin and his colleague Professor Einar
Hille. With the appearance of these publications Wiener was recognized
as the master in these fields, and was awarded the Bocher prize of the
American Mathematical Society in 1933. 5 In April 1934, he was elected
a Fellow of the National Academy of Sciences of the USA.
The mathematical significance of the memoir [32a] is commented
on extensively in the Wiener Call. Works, II. Here only three remarks
should suffice. (1) Tn 1938 H.R. Pitt from England proved a beautiful
theorem involving the so-called "slowly oscillating functions" of J.
Kamarta, a Yugoslav mathematician, from which both of Wiener's
general Tauberian theorems can be deduced. (2) The deep significance
of the methods employed by Wiener to establish his Tauberian theorems
came to light only eight to ten years later when the eminent Soviet
mathematician, I.M. Gelfand, developed the theory of Banach algebras.
(3) The further pursuit of Wiener's ideas at the abstract level gave birth
to an important branch of analysis called spectral synthesis. (Regarding
(2) and (3), see Note 3, p. 113).
While Wiener's memoir [32a] had been fully assimilated and
appreciated by the mathematical community, the same cannot be said
of his memoir [30a] on G. H. A. As late as 1970, the authorities E. Hewitt
and K. Ross remarked that it "has yet to be interpreted and assimilated
in the standard repertory of abstract harmonic analysis" {H7, vol. II,
p. 547}. Since this issue is thoroughly discussed in the Call. Works, II,
it will suffice to say that this writer has recently shown how [30a] can be
interpreted in terms of current functional analysis by means of helices in
Hilbert space. Helices, which are like the grooves on a cylindrical screw,
are familiar curves in elementary differential geometry, but the first helix
in an infinite dimensional Hilbert space appeared in Wiener's 1923 work
on Brownian motion. Thus this recent development affects an even
tighter integration of Wiener's G.H.A. and his theory of the Brownian
movement, and exemplifies the fundamental unity in this thought. (See
Note 4, p. 114.)
5 Named after the same Bocher with whose children he had played as a
six-year-old. Sharing the prize with him was Marston Morse, a former
student of G.D. Birkhoff. and another great mathematician-to-be.
109
lim - 1
2T
T-->oo
IT
-T
If(t) 12 a-f:,
cf. (3),
(21)
It
t-T
f(r)dr.
!T IIt-T
I E(r) I
2dr.
E(t)u,
110
lim -1
T .... oo
st
t-T
IE(r) I 2dr.
T .... oo
where !p( . ) is the covariance function of E( . ), thus reaching the conclusion that the intensity 7; recorded by the photometer is !p(O). Wiener's
replacement is valid if, in view of the chaotic origins of light, we make
the reasonable hypothesis (as Wiener did later) that E ( . ) is a trajectory
of a stationary stochastic process. For then the Birkhoff Ergodic Theorem (see 12A) tells us that the last two time-averages are indeed equal
in general.
Next, Wiener considered the Michelson interferometer, equipped
at the output end with a photometer: camera, eye, etc. A simple calculation shows that with the incoming light, represented by the electric
vector E( . ), the observed intensity at the output end is
!p(O)
!p(llt/c),
where Ilt is the difference between the lengths of the two arms, c is the
speed of light, and !p is the covariance of the signal E( . ). By turning the
screws, i.e. changing f1t, we can find !p(x), for any given (not too large)
x. Thus Wiener saw in the Michelson interferometer an analogue computer for the covariance function of light signals. It was his favorite
optical instrument, and he enjoyed exposing its principles to willing (and
even unwilling) listeners.
Wiener's explication of the notion of coherence of signals came
from his extension of G.H.A. to vector-valued signalsf = ifl' ... ,fq)
on IR. Wiener assumed that for each pairfi,fj, a cross-covariance function
!p if:
+ r)Jj(r)dr
(22)
S:oo eitAdF()'),
(23)
111
112
[Lh{j}](S)
Let s
f(s
h),
Lh
s E 1R3.
h)
roo
W(t-r)drx(r,o:),
0:
[0,1],
WE L 2 (R),
of the Brownian motion x(',), and showed that for almost all
IX, f( . ,IX) is in S and that for f( . ,0:) the spectral distribution F is
absolutely continuous with peA) = V(2n) I W( . ) I 2, a.e. where Wis the
indirect Fourier-Plancherel transform of W, [30a, 13].
The concepts of covariance qJ of a signal f and its spectral
distribution F, and their interconnection have an interesting history. The
first one to introduce F via its density F' was Schuster {S5} in 1899. But
he did so not by (14) but by the formula
P(J,) = lim _1
T~oo 2T
I-v
1
(2n)
f(t)e-iIAdt 12
(25)
-T
113
9 Notes
periodicities" in the irregular functions f that he encountered in empirical time-series. Schuster bypassed the covariance function rp, and was
probably unaware of its existence.
The fact that Einstein was a participant in this history was
revealed only in 1985, when a remarkable two-page heuristic note of his,
written in 1914, dealing with!, rp, F', and carrying a version of(13), came
to light {E3}. Einstein was unaware of the work of Schuster, and was the
first to bring in both rp and F' and reveal their linkage. Einstein's signal
f is on [0, 00], not II\t He defined F' essentially as had Schuster, but
worked with the "periods" (or "wave lengths") = l/A instead of the
"frequencies" A. He linked the function I(e) . - F'(l/e) to e by the
equation
I(e) =
2n
All the work was done independently. A detailed description of Einstein's paper and its nexus with G.H.A. can be found in the writer's 1986
article {M12}. An interesting stochastic process interpretation of Einstein's note, and its links to Khinchine's work of 1934, has been given
by A.M. Yaglom {Y2} in 1987. He also gives the facts of the discovery
of Einstein's paper by a school teacher in East Germany, and other
interesting historical information.
Note 3: Spectral synthesis (cf. 9E).
114
of], then the last result is restatable in the following form, which Wiener
might have found even more esoteric: if Z(]) is void, then the smallest
closed ideal containing f is equal to the intersection of maximal ideals
containing f Does the italicized conclusion prevail when Z(]) is nonvoid? Efforts begun in the late 1930s to answer this question culminated
in a negative answer by P. Malliavan in France in 1959. These efforts
have brought into being a whole new branch of mathematical analysis
called spectral synthesis, with distinguished participants all over the
globe. Its development is traced systematically in the treatises of Hewitt
and Ross {H7, vol. II}, and J. Benedetto {B6}, and their accounts show
how Wiener's ideas, in one form or another, keep lurking in the background.
Note 4: Conditional Banach spaces (cf. 9E).
From the standpoint of functional analysis, the Wiener class S is a
conditionally linear subspace of the Marcinkiewicz Banach space:
mill\)
{f:jEL\oC(II\)&llfIIZ
(Xl},
(cf. Coli. Works, II, pp. 333-379). K. S. Lau {L2} has shown that the f
in S with I f I = 1 are in fact extreme points of the unit ball of mill\).
Recently, he has obtained an interesting extension of BMO spaces by
considering the class in which the Marcinkiewicz lim is replaced by sup,
cf. {L3}.
Wiener's generalized Fourier transform s of the function fin S,
cf. (l5), gives rise to a helix in the Hilbert space L2(1I\). Every such helix
is characterized by a single 'average vector' IX in Lz(II\). This IX turns out
to be the Planche rei transform off- to, where to(t) = l/(t + i) Vn, and
the Tauberian identity (17) holds with ex replacing Wiener's s. Thus there
is a simpler theory, accomplishing all of Wiener's objectives, but with
the generalized Fourier transform in the Hilbert space L2(1I\).
At the more classical level, J. Benedetto has just obtained a
full-fledged generalization of the Tauberian identity (17) for functions f
in II\n in the Wiener class S, cf. J. Benedetto, G. Benke and W. Evans,
"An n-dimensional Wiener-Plancherel formula", Adv. Appl. Math.
(forthcoming).
115
roo
tf(t) 2dt
1
roo
Aj(J,) 12dA
~~,
(1)
Herefis in L 2 (1R), with norm 1, i.e. SCXJ oo If(t) I 2dt = I, and the integrals
e'oo I Aj(A) I 2dA are finite, j being the (indirect)
Fourier-Planchere1 transform off
It follows that if a sound oscillation f of intensity (or loudness)
1 lasts only for a short time interval [0, 7], then the first factor on the
left will be small; consequently the second factor will have to be large,
i.e. the oscillationfwill have to comprise a whole range of frequencies
A, and will not be a pure tone. Conversely, if the oscillation approximates a pure tone of frequency Ao, i. e. its frequencies are all clustered around Ao, then the second factor will be small; the first factor will
now have to be large, i.e. the oscillation is spread over a long interval
of time. In short, a pure tone of only momentary duration is impossible.
SOOoo I tf(t) I 2dt and
116
10
Wiener enjoyed reflecting on the practical consequences that this entailed in acoustics and music. In
musical notation for instance, the vertical position on the
staff indicates the pitch, i.e. frequency, and the horizonlit
tal interval the metronome duration of a full noteo, say
",---,''----1
112 seconds. The quarter note will only be 1/8th of a
second in duration. As this note is moved down the staff
(see Fig.), its frequency diminishes. Were this quarter
note (lasting 1/8th of a second) to be played on some
instrument, at the very low frequency of 5 oscillations per
second, not even one oscillation will be completed, and
the air will be pushed, not set into vibration. What we
will hear will sound not like a note "but rather like a blow
on the eardrum" [56g, p. 106].
Thus for a composition to be musical, its notes must throughout
satisfy an inequality of the type:
117
T~oo
118
10
The role of the Born-Wiener paper [26d] in the history of quantum mechanics is alluded to in Whittaker's history {WI2, Vol. II,
p. 267}, and discussed much more fully in J. Mehra and H. Rechenberg's
recent comprehensive history of the subject {M 15, Ch. 5}. The idea of a
Hilbert space, still embryonic, was just not in Wiener's consciousness at
that time. But the paper [26d], its limitations notwithstanding, had "an
immediate impact on Heisenberg", as Mehra and Rechenberg point out
{MIS, p. 247}, and they conclude:
At a time when just a few physicists struggled to develop a consistent theory
of quantum mechanics, the Born-Wiener collaboration not only indicated
the way for handling the problem of aperiodic motion but also contributed
to the physical interpretation of the theory. {MIS, p. 246}
119
J(t)
I.~ akeiht,
then
!pet)
I.~ I a k 12eiAkt.
The phases of the complex numbers a k are gone. Feeding this "observed
light" into another optical instrument will not produce the same response as feeding in the unobserved lightf Thus observation affects the
signal and thwarts prediction, much as in quantum mechanics. If we
think of the light beam as a vector-signal, then in Wiener's words:
... if two optical instruments are arranged in series, the taking of a reading
from the first will involve the interposition of a ground-glass screen or
photographic plate between the two, and such a plate will destroy the phase
relations of the coherency matrix of the emitted light, replacing it by the
diagonal matrix with the same diagonal terms. Thus the observation of the
output of the first instrument alters the output of the second. [30a, p. 194]
120
10
121
Wiener gave an epistemological interpretation to Leibniz's principle of the identity of indiscernibles: what cannot be observed cannot
be discerned and must be abandoned. Thus for him, Leibniz was the
spiritual grandfather of the operational method brought into the theories of relativity and quanta by Einstein and Heisenberg and advocated
as a general methodology by the physicist P. W. Bridgman {B20},
Wiener's co-student at Harvard. Wiener also referred to the central
position of the principle of least action in modern physics, and tracing
its origin back to Maupertius, suggested that it follows from Leibniz's
122
10
123
More must be said about Haldane's paper because of its influence on Wiener and its place in the development of cybernetics.
Living organisms self-regulate, self-repair and self-reproduce, i. e. act as
organic units in seeming defiance of the laws of physics and chemistry.
To incorporate the living organism into science holistic postulates of one
sort or another had been proposed by vitalists. Haldane points out that
the atoms of physics also exhibit an organic unity. For instance, they
quickly "repair" the loss of an electron by picking up another-a behavior that the Rutherford, Bohr theories of the atom in the 1910s could
not explain. Physicists healed this breach not by adding holistic postulates to these theories, however, but by discovering wave-mechanics (c. 1923). The de Broglie, Schrodinger theory was able not only
to account for the observed atomic behavior, but to reveal the existence
(then unknown) of two isomers of the hydrogen molecule. In his paper
Haldane claims that this wave-mechanics serves to explain biological
and psychological facts that were a mystery from the standpoint of
classical physics and chemistry. The last forty years have borne out
Haldane's claim remarkably well. His paper is a fine example of the
potency of bold speculation when it comes from a learned and imaginative mind.
Haldane singles out the following deviations of the wave mechanical particle from the classical particle as especially germane:
A) The electron deviates from the classical particle not only in
regard to the simultaneous indeterminability of its position and momentum, but in its loss of identity during interaction. Two electrons are
extrinsically distinguishable, for instance by their different locations at
a given moment. But their strange behavior when they approach one
another precludes our keeping track of which is which in the future. In
fact, the probability that the two exchange positions tends to 1 as their
distance apart tends to O. The expression "this electron" is therefore
meaningless. Likewise for the helium nucleus, the so-called IX particle.
B) For these particles the probability of spontaneously crossing
("leaking") over a potential barrier (i. e. attaining states of higher potential energy) is positive, as the phenomenon of radio activity reveals. 2
For the classical particle, such a cross-over is impossible without the
supply of outside energy.
C) The stable quantum mechanical configurations-atoms,
molecules, crystals-are states in which the probability of exchange is
high, i.e. states of so-called high internal resonance, or high degeneracy
in the sense of loss of degrees of freedom:
2 This quantum mechanical property is put to use in the tunneling transistors
now in vogue.
124
10
125
126
10
1646-1716
127
J.B.S. Haldane
(c. 1930)
1892-1964
128
that Niels Bohr describes in his book {BI4}. In the course of this paper
Wiener describes science as a technique for wading through a sea of
pointer readings: "Physics is merely a coherent way of describing the
readings of physical instruments" [36g, p. 311]. This half-truth, clearly
at variance with his earlier views, was soon to be abandoned. It is one
of the solitary instances in which this very coherent thinker articulated
badly.
129
I PIfII~,
2.
(2)
2.
(I)
In this representation, the choice of the space (Q, d, p) and the mapping
Mp are crucial. It will suffice to say that (Q, d, p) is the space of a
multiparameter Brownian motion (Ch.7, Note 5). In this way the
Brownian motion enters the picture. For more on this and on M p , see
the Note, p. 130.
By establishing (1) Wiener and Siegel fulfill their goal of showing
that the probability fl(P) appearing in quantum mechanics as the square
of the amplitude I PIfII yt' is the probability of a well-defined subset
M /P) of a well-understood probability space (Q, d, p), viz. a
generalized Wiener space.
Unfortunately, it is not clear from the Wiener-Siegel work if an
equality of the type (I) is available for impure states fl. Furthermore,
Wiener and Siegel do not discuss the physical relevance of the Brownian
motion as yielding "hidden parameters" in quantum mechanics,
although this notion is central to their approach. This research is unfinished, and its significance is as yet far from clear.
The possibility that Brownian motion theory may have a role to
play in quantum theory has occurred to several scientists, e. g. Bohm, de
Broglie and Vigier, who follow Einstein and question the completeness
of the existing quantum mechanics, and believe that the hiddenparameter problem is worth investigation, but perhaps in formulations
differing from von Neumann's. As de Broglie has put it in his foreword
to Bohm's book:
To try to stop all attempts to pass beyond the present viewpoint of quantum
physics could be very dangerous for the progress of science and would
furthermore be contrary to the lessons we may learn from the history of
science. This teaches us, in effect, that the actual state of our knowledge is
always provisional and that there must be, beyond what is actually known,
immense new regions to discover. Besides, quantum physics has found itself
for several years tackling problems which it has not been able to solve and
seems to have arrived at a dead end. This situation suggests strongly that an
effort to modify the framework of ideas in which quantum physics has
voluntarily wrapped itself would be valuable. {B 13, px}.
130
Involved in the crucial mapping Mil is the homogeneous chaos corresponding to the complex valued q-parametrized Brownian motion defined in Ch. 7, Note 5. (For the homogeneous chaos, see 12E below.)
Following Kakutani {K3} we may introduce this chaos, also called the
Brownian motion random measure by the four conditions:
(a) Its domain is the 6-ring q; q of all Borel subsets D of IRq for which
m(D) < 00, m being the q-dimensional Lebesgue measure;
131
10 Note
(b) for each D in f!) q' ~(D) is a real- or complex-valued random variable
over a probability space (Q, d, p), which is normally distributed
with mean 0 and variance m(D); 3
(c) ~ is countably additive under the norm of the space L 2Q, d, p);
(d) ~ is independently scattered, i.e. for disjoint D 1, , Dn in f!)q, the
random variables ~(D I), ... , ~(D n) are independent.
An elegant definition can be given for the stochastic integral
{Ju;Jq !p(x)~(dx)}('),
where !p
LilR~,
(3)
WE Q,
the last being the stochastic integral mentioned in (3). Next, define
MjP) = {w: WE Q & IfJl(P)(w) I 2: If/P~)(w) I}
IPIfIIL(u;J') = f1(P).
Thus the equality (I) is established. Note that we can write (I) in the form
f1(P) =
QVJl(w)(P)p(dw)
or
f1 =
QvJl(w)p(dw),
(II)
3 The normal distribution N over IC with mean 0 and standard deviation (J (i. e.
variance (J2) is defined by
N(B) =
1
(JV(2n) e
1 Z 12
/2(J2
v(dz),
Bc:IC,
where v is the Haar measure over IC, which assigns the value I to the unit
square in IC. Kakutani, like Wiener, considered only real valued random
variables ~(D). But the complex case is important, particularly for the
Wiener-Siegel work on quantum theory.
132
11
133
Eberhard Hopf
1902-1983
S~ W (t - r)J(r)dr,
t;:::: 0,
from which the unknownJ can be found. This is the homogeneous case
of the general equation
with g =
J(t) = get)
+ S~ K(t -
s)J(s)ds,
t;:::: 0
(1)
where K( ) and g( . ) are given, and the unknown isf( . ). The eminent
astronomer E. F. Freundlich had apprised Hopf of the need for a mathematically more mature treatment than the one known to the astronomers. Wiener and Hopf proceeded to give such a treatment for the
homogeneous case of (1).
134
11
Hopfs collaboration with Wiener was very short. They discussed the equation intensely one afternoon in Wiener's New Hampshire home. Wiener, who had come across similar equations in his study
of electric circuits, saw the causal significance of the equation when t is
interpreted as time. (In the radiation problem t is the optical depth, as
we saw.) The next morning he came up with a solution. This turned out
to be wrong, but the methods which he had employed proved extremely
fruitful. Hopf was able to clear up the error and obtain a complete
solution without much trouble. 2 The result was their famous joint
paper [3la] Uber eine Klasse singuliiren Integralgleichungen. Wiener
wrote only one other paper on the subject [46a] in collaboration with
A.E. Heins.
B Significance of the Hopf-Wiener equation; causality and analyticity.
It soon became clear that equations of the Hopf-Wiener type have wide
applicability, as they cover many situations in which a barrier separates
two different regimes, one of which can influence the other but not vice
versa. In stellar radiation the barrier is the surface of the radiating core,
and conditions inside influence those outside. Likewise with the atomic
bomb, which is essentially the model of a star in which the surface of the
bomb marks a change between an inner regime and an outer regime. But
Wiener also perceived that in problems of Bergsonian temporal development, the "present" acts as a buffer between the influencing "past" and
the indeterminate "future". The cogency of this perception emerged
about ten years later when Wiener's war work on anti-aircraft fire
controlled him to his theories of prediction and filtering, and he readily
came up with a Hopf-Wiener equation from his data. This equation,
however, is of the form
get)
S:
K(t - s)f(s)ds,
t;:::: 0
(2)
get)
S~f(t - s)W(s)ds,
IR
135
(3)
which is akin to (2). But the complete lack of restriction on tin (3) makes
it much easier to deal with than the equation (2).
Wiener brought many fine ideas to bear on this border area of analysis,
but their fruition demanded detailed work for which he did not always
136
11
137
R.E.A.C. Paley
1907-1933
33c, 33e] and in the American Mathematical Society Colloquium Lectures that Wiener delivered in 1934, and which he published under their
joint names [34d]. This book, which covers a great many topics, is
"unified by the central idea of the application of the Fourier transform
in the complex domain". It demonstrated that a wide range of problems
from complex function theory are amenable to a treatment by Fourier
transformation. It was a very important contribution to the border area
we spoke of, in which the methodologies of real and complex analysis
have to be fused. This area continues to occupy a central position in
research, and even advanced pedagogy, cf. Rudin {R3}.
Paley's untimely death came as a severe blow to Wiener. On him
fell the sad duty of notifying Paley's mother and his British friends. "It
took me some time to come back to a mental equilibrium sufficient to
permit my further work ... " [56g, p. 170]. Some of Wiener's words
about his collaboration with Paley are worth quoting, since they reveal
both his colossal indebtedness to Paley, as well as the difference between
138
II
11/(1 +
Je 2)dJe <
00.
139
12
Wiener also valued the Individual Ergodic Theorem for the new light it
shed on his generalized harmonic analysis. The result that any strictly
stationary stochastic process is equivalent to one governed by a measure-preserving flow, cf. {DS, pp. S09-S11}, allows us to re-enunciate the
theorem in Wienerian terms:
Ergodic Theorem (G.D. Birkhoff, 1931). Letlf(t,'): - 00 < t <
oo} be a complex-valued strictly stationary stochastic process over the
probability space (Q, flll, P), and let 1(0, .) E L 2 (Q, flll, P). Then for P
almost all OJ in Q, the signal f( . , OJ) belongs to the Wiener class S, and
its covariance function rp( . , OJ) satisfies the equality
140
12
T E IR
(1)
T E IR,
(2)
Wiener valued the ergodic theorem not just for its justification of
the deep equalities (1) and (2), but also for its validation of his longcherished belief that
lim -I
T
T~oo
JO
-T
f(t
+ T)f(t)dt
lim - 1
2T
T~0U
JT
-T
f(t
+ T)f(t)dt.
(3)
The average on the left involves only the past values of the signalf and
is thus in principle estimable from observation, whereas the average on
the right is not, since it involves the unknown future.
The Birkhoff theorem thus gave Wiener just what the Gibbsian
in him wanted to know: past observations of the signal suffice for the
estimation of its covariance function, and this function gives information
intrinsic to the stochastic process from which the signal hails. The fact,
that ergodicity has to be postulated in order that the covariance may
have an unconditional probabilistic significance, did not deter Wiener at
all. This assumption was just one of many idealizations that his scientific
philosophy permitted. Moreover, von Neumann's celebrated 1932 theorem on the disintegration of regular measure-preserving flows over
complete metric spaces into ergodic sections {V4} gave Wiener a good
excuse to deal almost exclusively with the ergodic case, cf. for instance,
[61c, pp. 55-56].
t, a) - x(a
t, a)
= x(b,
IR,
(4)
141
hm T~oo 2T
S
1
._
f(t,OJ)e-Wdt
exists.
-T
They also gave conditions under which such a signal f( . , OJ) will be a
Besicovitch almost periodic function. These results are significant in
view of the presence of lots of functions in the unrestricted Wiener class
S that are not cross-correlated with the characters e AIn an earlier paper [38a] Wiener embarked on an even greater
generalization of the ergodic theorem. With an eye on the intractable
problem of turbulence, he brought ergodic concepts to bear on the
stochastic integrals of the Brownian motion which he had introduced
much earlier. The inception of these ideas depended, however, on the
much clearer perception he had by then acquired on the nature of cosmic
contingency. To this we must now turn.
C The contingent cosmos, noise, and Gibbsian statistical mechanics
The understanding that Wiener acquired from his ergodic work was
instrumental in deepening his earlier appreciation of Gibbsian mechanics.
Gibbsian mechanics is an offshoot of classical mechanics. In the
latter we assume that the material system placed in a field of force is
deterministic, i.e. given the laws governing the field of force and the
142
12
[R2N,
I .... , N,
where H is the function on.ff x [R such that H(q, P, t) is the total energy
of the system in the phase (q, p) at the instant t. H is called the
Hamiltonian of the system. (For its definition see Note 1, p. 157).
It follows that for time-independent force-fields, H cannot depend on t. It also follows that the field is conservative, i. e. the total
energy of the system does not change as it moves. Consequently, the
phase trajectory of any system with energy Eo is confined to the
Eo-energy hypersurface
L: H(q, p)
Eo,
This is because at best the molecules are rigid, and a rigid body has 6 degrees
of freedom. Its configuration is determined by fixing three non-collinear
points A, B, C, which we may suppose are at unit distance apart. This gives
9 coordinates. But these are not independent since AB = BC = CA = I. The
number of independent coordinates is 6.
143
It was Gibbs's key idea to define the "state" of such systems for
which N is enormous not by the phase (q, p), but by a probability
distribution P over the phase space. Now the state-space is the class of
probability distributions P over the phase space. The problem of statistical mechanics is to deduce the evolution of the probability P, given the
laws governing the force-field and the probability Po at some initial
moment to.
Even with a deterministic cosmos the Gibbsian attitude of bringing in probabilities has considerable scientific merit. But it assumes a
compelling inevitibility when judged from the perspective of the molecular kinetic conception of the world that emerged from the efforts of
Maxwell and Boltzmann in statistical mechanics. As we saw in 7D, this
perspective was accepted without skepticism after the confirmation of
Einstein's theory of the Brownian motion in the 191Os.
This molecular kinetic theory of matter has a pre-history that
goes back to Democritus. It tells us that all atoms perform at all times
a disorderly motion, which is, so-to-speak, superposed on their more
orderly movement. This chaos does not allow the happenings between
a small number of atoms to involve themselves according to any recognizable laws. Only in the cooperation of an enormously large number of
atoms do statistical laws begin to operate, their accuracy increasing with
the number of atoms. All the supposedly causal laws of physics and
chemistry that operate at the macroscopic level are really of this statistical kind. At the microscopic level, however, we have patently probabilistic laws. These govern the kind of random phenomena of which the
Brownian motion and the shot effect are good examples.
Random phenomena, like the Brownian motion, which occur at
the microscopic level, tend to disturb and dampen the systematic movement of matter and energy due to a definite stimulus. We therefore
describe them as noise when we wish to refer to their disturbing role. For
instance, the messages (electromagnetic pulses) transmitted over the
radio are disturbed by atmospheric noise. The messages (chemical reactions) flowing across a nerve in the human body are disturbed by noise
caused by sporadic movements of particles within the body, and so on.
The systematic movement of mass and energy in an instrument,
(e.g. a chemical balance) due to a definite stimulus (e.g. a specimen in
one pan and weights in the other) is likewise disturbed by noise, and this
affects the equilibrium-position of the pointer. This explains why when
we perform repeated trials of a controlled experiment, our results keep
fluctuating. To reduce the effects of noise we will have to make the
instrument insensitive, thereby diminishing its accuracy and utility. To
make it accurate, we will have to increase its sensitivity, but then it will
begin to register noise along with the message, and again lose its useful-
144
12
J.-B. Fourier
1768-1830
H. Lebesgue
1875-1941
145
l.W. Gibbs
1839-1903
G.D. Birkhoff
1884-1944
(c. 1930)
146
12
ness. Clearly, a perfectly accurate and universally useful measuring instrument is impossible, even in principle.
Such examples, and the intrinsically stochastic aspects of subatomic phenomena, suggested by quantum theory (Ch. 10) (unknown to
Maxwell, Boltzmann and Gibbs) force us to recognize that there is a
chance or random element in the very texture of Nature, and that the
orderliness of the world is incomplete. For scientific purposes we can no
longer regard the universe as a strictly deterministic system, the phase of
which at any instant t is determined exclusively by its phases at all
previous instances t' < t. It was Wiener's position that the universe is
still a cosmos, and that the Principle of the Uniformity of Nature still
reigns. But it reigns at a stochastic level: it is the probability measures,
engendered by the statistical aspect of Nature, that remain invariant
under time-translations. (See [50j, Introd.] and [55a, pp. 251-252].)
Wiener held that Gibbs's idea of defining the "state" of a system
as a probability distribution is indispensable in view of the contingency
of the cosmos. This belief came from his realization that our observations of such systems are confined to a few macroscopic parameters,
such as pressure or temperature, and that even their measurements are
inevitably inaccurate because of the random impacts of Brownian motion on our instruments, especially the more sensitive ones. But a study
commencing with a probability distribution can only yield stochastic
quantities: new distributions, expectations, variances and the like. On
the other hand, our observations on the system can only tell us how it
behaves over different time-intervals. To fulfill Gibbs's program, i.e. to
be able to judge the hypothetical probability distributions P empirically,
some assurance is needed that the time-averages are equal to the corresponding phase-averages. This assurance the ergodic hypothesis was
supposed to provide.
This line of thought began with Maxwell and Boltzmann, who
viewed the gas enclosed in an insulating vessel (with perfectly adiabatic,
perfectly rigid and infinitely massive walls) as a conservative system with
a very large number N of degrees of freedom. The phase space is now
[R2N, and the orbit of the gas is confined to the energy hyper surface
L:
H(q, p) = constant,
T~oo
IT 1 (T,x)dt,
R
x EL,
147
148
justifies Carnot's principle that the efficiency of all steam engines is less
than one. It enables us to grade energy into high, medium and low
grades. It makes available the concept of the (Helmholtz) free energy,
viz. U - TS, i. e. energy ideally convertible into useful work, a concept
of considerable ecological and technological importance. Here S is the
entropy, T the temperature we spoke of, and U the internal energy, cf.
Note 2, p. 157.
Classical thermodynamics gives way to statistical thermodynamics when its concepts are interpreted in terms of the molecular kinetic
theory. From the standpoint of the latter, a thermodynamic system rt. is
an assembly of a very large number of atoms, which perform at all times
a completely disorderly motion, in addition to any orderly movement
they may acquire from the action of a definite external field of force. The
mechanical energy of this chaotic movement is the quantity of heat in the
system. The mean kinetic energy of the random motion of the atoms is
its absolute temperature. Many familiar thermodynamic facts receive a
very natural explanation in these molecular statistical terms. One instance is the flow of heat from a body at high temperature to one at low
temperature; another is Avogadro's Law to the effect that all gases
having the same volume, pressure and temperature have the same
number of molecules. But many other facts, incomprehensible from the
classical thermodynamic angle, such as the Brownian motion, are also
explainable.
Especially relevant to Wiener's work is the statistical interpretation of the thermodynamic entropy S, or briefly S, that Boltzmann
proposed in the 1870s. Consider a mass of dilute gas consisting of n
identical molecules, each of r degrees of freedom, enclosed in a cylinder
with piston. The molecular phase space Q is contained in 1R(21, and the
gas phase space Qn in 1R(2rn. An observable macrostate s of the gas, (i.e.
a thermodynamic state such as the one in which its volume is v and its
pressure is p) will comprise an enormous number of phases of the gas,
and these will form a subregion s* of Qn having a large 2rn-dimensional
volume V(s*).
By the Ergodic Theorem it turns out that the mean-time of stay
of the gas in the macrostate s depends on this volume V(s*), cf. 12A.
Boltzmann defined the statistical entropy of the gas in the macros tate s
in essence (apart from unimportant constants) by the equation
Ent(s)
k log V(s*),
(5)
Note 3, p. 158.
Boltzmann interpreted Ent(s) as a measure of the internal disorder of the thermodynamic system in state s. For many systems it can be
149
shown that Ent(s2) > Ent(s)) when S(S2) > SCSI)' S being the thermodynamic entropy. This authenticates Ent as the statistical correlate of
S. Boltzmann attempted to derive the law of increasing entropy by
showing that
dEnt(s,)/dt 2 0,
150
12
l1(a, b](ct)
ct
[0, 1]
(6)
!J
lim _1_
HOG
where vCr)
(t), .. . ,t,J
( absolute)
transitive"
vCr)
t) (w)}dt
exists,
tl
f{I1(A
V(r)
G &
I1(A
t)(w)
H}
P{w: I1(A)(w)
H},
~oo
The measure 11 given by (6) is not merely stationary: it is independently scattered, i.e.l1(a, b) and I1(C, d) are stochastically independent for
disjoint intervals (a, b) and (c, d). Wiener called this 11 the I-dimensional
pure chaos. Repeated stochastic integration (cf. 7D) with respect to this
11 yield derived or polynomial chaoses. For instance, over 1R3 we get a
polynomial chaos v given by
v(B) (ct)
HJ
for any Kin L2 (1R3); here B S; 1R3 and ct E [0,1]. Wiener's theorem [38a,
12] is that every ergodic homogeneous chaos 11 over IRn can be "approximated" (in a reasonable sense) by a sequence Ill' 112, ... , 11k> ... of
polynomial chaoses over IRn.
Wiener's hope was to bring these considerations to bear on
statistical mechanics. What he cherished may be summed up very roughly as follows. In continuum statistical mechanics (e. g. in the statistical
theory of fluids or plasmas) the initial distubance fo is a random phase
field, and the dynamical transformation U, such that U,(fo) is the
disturbance at instant t, is ergodic. The ergodic action is well understood
when it is applied to the I-dimensional pure chaos 11 of (6) or to one of
151
the chaoses v derived from it; from 12B(4) we get the equality, Ut{v(B)}
= v(B + t). Consequently, iffo can be expanded in a series of multiple
stochastic integrals with respect to such chaoses, we get a handle on
finding Ut(fo).
In subsequent work which appeared in [58i], nearly 20 years
later, Wiener showed that a large class of functions f admit such series
representation. In this work he was guided at least indirectly by the
important contributions of Cameron and Martin {CI} and Kakutani
{K3}. While the purely mathematical value of this work initiated in [38a]
and in a cognate paper with A. Wintner [43a] on the discrete chaos, is
beyond question, its import in statistical mechanics is not yet clear. This
issue is discussed fully in the survey of Drs. McMillan and Deem in the
Call. Works, I, pp. 654-671. Wiener was well aware of the difficulties in
his approach, and clearly pointed them out in [38a]:
In many cases, such as that of turbulence, the demands of chaos theory go
considerably beyond the best knowledge of the present day. The difficulty is
often both mathematical and physical. The mathematical theory may lead
inevitably to a catastrophe beyond which there is no continuation, either
because it is not the adequate presentation of the physical facts; or because
after the catastrophe the physical system continues to develop in a manner
not adequately provided for in a mathematical formulation which is adequate up to the occurrence of the catastrophe; or lastly, because the catastrophe does really occur physically, and the system really has no subsequent
history. [38a, p. 935]
152
12
153
The more unlikely the outcome x, the smaller will be the value p{x}, and
the greater the value log[ljp{x}], i.e. lex). Thus the definition meets the
reasonable criterion that the more a message removes uncertainty, the
greater its informative value. 3
For the simplest p, viz. one with two outcomes of equal probability Ij2, the message that the first (or second) of these outcomes has
occurred has the value log 2. It is convenient to define this to be unit
information, and accordingly to take 2 as the base for the logarithm. To
transmit this information, it suffices to flash just one of the digits 0 or
1 according as the first or second alternative has occurred, i. e. flash
exactly one binary digit. The unit is therefore called the hit (for "binary
digit"). If P happens to have exactly 2 n atomic outcomes, then the
3 Cognate ideas appeared earlier in the maximum likelihood method of the
statistician R. A. Fisher, c. 1920.
154
12
E[ -log p{ .}] = - I
(7)
XEX
We may call this the average informative value of the probability distribution p. It gives the average energy and (average cost) of transmission.
This definition, in which X is finite or at most countable, is due
to Dr. C.E. Shannon in 1947 or 1948 {S7}, who was then at the Bell
Telephone Laboratories and was concerned with the energy-efficient
coding of telegraphic messages over noisy channels. It presaged his deep
work on channel capacities, encoding and decoding.
In the summer of 1947 Wiener was led to the same problem for
an absolutely continuous probability distribution p over the real line IR
by the needs of filter theory. Since an infinite sequence of binary digits
is required to transmit a real number, a limiting approach, starting with
the Shannon concept, will assign an infinite average information to such
a p. Wiener's starting point was the observation that we may forget all
digits after a fixed number because of noise. By an argument, very
obscurely presented, he arrived at the following definition for the
average informative value of p:
Inf(p)
E{ -log p'( . )}
(8)
where p' ( . ) is the probability density, cf. [61c, p. 62, (3.05)].4 He too
took logarithms to the base 2. This definition is widely used in the theory
of information processing for continuous time.
Wiener's Inf. unlike Shannon's, can become negative. Nevertheless both Infs have essential common features. For instance, both Infs
are additive for independent information-sources. We may therefore
speak of a single Shannon- Wiener conception of average informative
value. It provides the foundation of a large and important branch of
contemporary engineering.
The nexus between communications engineering and statistical
mechanics, which Wiener had dimly discerned in the mid-1920s,
4 Wiener omitted the minus sign, but this is of little consequence since Inf(p)
can take either positive or negative values for different p.
155
C.E. Shannon
(cf. 9C), is deep indeed, for the Shannon-Wiener concept of information has turned out to be a disguised version of the very concept of
statistical entropy to which Boltzmann was driven 70 years earlier. For
the proof, see Note 4, p. 158.) Let it also be noted that to prove his
H-theorem, dEnt(st)/dt ;::: 0, ( 12D), Boltzmann had to replace his
summation by an integral. This integral is precisely the Wiener average
information of a multidimensional probability distribution, cf. [6lc,
p. 63] and Born {BI7, pp. 57 (6.25), l65}.
Whereas the Boltzmann entropy of a gas is best interpreted as a
measure of "internal disorder", Shannon's average information is most
naturally interpreted as a measure of "uncertainty removed". Their
equality has suggested the term negentropy, or measure of internal order,
as a substitute for the term information in certain contexts.
The cogency of this viewpoint has become clear from the work
of Szilard, Brillouin and also Wiener [50g, 52aj on the Maxwell demon.
Dr. 1.R. Pierce {P5, pp. 198-200, 290} defines this demon as
156
12
157
12 Notes
T(q, p),
where
N N
T(q, p)
L L a;j(q)q;pj.
;=Ij=!
is the kinetic energy of the system in the phase (q,p). Here aif(q) are the
coefficients of inertia; they form an NxN non-negative symmetric matrix
[aif(q)]
dS a ( )
on 9',
whereS a(' ) is a function on 9'. The function T{O J. )} on 9' is called the
absolutetemperature,andS a ( ' ) the corresponding entropy of the system.
The function T() on IR is not entirely unique. The simple
determination T(t) = c' t, C = const., suffices, provided that the
empirical temperature e,(s) of a in the state s is measured by the "perfect
gas thermometer". (This is defined as the thermometer which assigns to
a perfect gas in the state (p, v) the temperature pv / R. In the laboratory,
hydrogen is used as the "perfect gas"). With this standardization of
T( . ), it can be demonstrated that the entropy S J . ) of the system a
cannot decrease in any adiabatic transformation, and must increase in all
non-quasi-static adiabatic ones.
158
12
The reason for the logarithm is to make the entropy additive for noninteracting systems. If s = (s I, s 2) is the macrostate of the mixture of two
such systems (e.g. gases with Sl = (v 1,PI)ands 2 = (v 2,P2),thens* =
s; x
whence, on taking the appropriate hypervolumes, we get
s;,
and
Ent(s)
To see this, recall from 12D(5), Boltzmann's expression for the entropy
of a thermodynamic system in a macrostate s, viz.
Ent(s) = k . log V(s*),
where V (s*) is the 2m-dimensional volume of the region s* in the gas
phase space Qn spanned by s. Notice if the volume of a subregion ,1 of
Qn is below a certain threshold, there will be no observable way of
knowing that the phase of the gas is in ,1. Accordingly, with Boltzmann,
let us partition the gas phase space Qn into a large number of cells ,1 of
equal volume, say Vo. Each of these cells is called a microstate of the gas.
Their collection is a finite partition II of Qn. For each macrostate s of
the gas, s* will comprise a large number of the microstates, i. e.
s*
U ,1,
A
(9)
TTs
It follows that
V(s*)
V(,1)
.dEn,
and that
Ent(s) = k 10g(Vo . card IIs)
(10)
159
12 Notes
and clearly Vo = Va. Let the numbers of molecules with phases in the
cell (J be Il.(a). Then the function Il. on n to {a, 1, ... ,n} is called a
complexion of the gas. Attached to each complexion Il. is a probability
measure P A( . ) over the set n given by
p "(a) = ll.(a)ln,
n.
U ,1,
where II}
S;
II .
.JEll,.
= n!1
II Il.(a)!.
Assuming that each /c(a) is large enough to allow Stirling's approximation to Il.(a)!, we find that
log{card II}.}
-n
~
aEn
Thus by (10),
Ent(ll.)
kn . log e2 . Inf(p A)
k . log Vo.
In words, the Boltzmann entropy of the gas in the complexion Il. is equal,
apart from an additive constant, to kn . log e2 times the Shannon average
information of the associated probability measure P A' on the partition n of
the molecular phase space. Here k is Boltzmann's constant and n is the
number of molecules.
5 It is assumed that each (J, though very small from our macroscopic standpoint, is very large from the atomic, i.e. an enormous number of molecules
will have phases lying in (J.
160
13
or
S:f(t - x)g(x)dx,
161
rule, consisting of two sliding rulers graduated according to the logarithms of numbers. (To multiply the numbers a and b, we need only
read off the graduation for the sum log a + log b.) Many of the fine
astronomical instruments designed by Islamic scientists of the 14th
century are analogue machines for specific computations. Lord Kelvin's
harmonic analyzer, to compute the first few Fourier coefficients of the
periodic rise and fall of tides at a fixed spot, is another example of an
analog computer.
Wiener was fascinated by Bush's program, and his close touch
with it significantly molded his thought. His interest in the fabrication
of thinking machines had begun in college when he studied Leibniz's
ideas of this question (cf. Ch. 4), and he had used computers during his
army service at the Aberdeen Proving Grounds (Ch.6). From the
engineers, Wiener now learned of pro blems involved in the designing of
workable computers. This was to contribute to his understanding of the
problems of automatization. For their part, the engineers, aware of the
service rendered to computer science by philosophical minds such as
Pascal and Leibniz, found in Wiener a useful person to have around. His
flair for electrical and mechanical devices, his practical computing experience, hs mathematical abilities and philosophical training were a
rare combination, and he was known to throw out interesting ideas on
all sorts of subjects, many of them worth a try.
Wiener's first concrete contribution came in his advice on the
continuous integraph. In Bush's 1927 model {B24}, a sheet of paper
bearing the graphs y = fix), y = g(x) was attached to a moving table.
Presiding over the table were two operators, each holding over one of
the graphs a pointer attached to a vertically movable slider of a potentiometer carrying a constant current. The ordinates of the graphs were
thus turned into voltage drops, and these were fed into the potential coil
and the current coil of a modified Watthour meter. By a relay mechanism, the reading of the meter was then continuously recorded on
another sheet of paper attached to the moving table. Thus as the table
moved, the stylus traced out the desired curve y = J~f(x)g(x)dx. (The
ordinary household Watthour meter performs the integration J~ v(t)i(t)dt,
where vet) and i(t) are the voltage and current, and thus gives the energy
consumed during the time-interval [0, 1].) This ingenious device fails to
deliver a single graph, however, when the parameter t occurs in the
integrand, as for instance in the integral J(t) = S:f(t - x)g(x)dx.
In early 1926 Wiener found a solution. This is how it happened:
One time when I was visiting the show at the Old Copley Theatre, an idea
came into my mind which simply distracted all my attention from the
performance. It was the notion of an optical computing machine for harmonic analysis. 1 had already learned not to disregard these stray ideas, no
162
13
matter when they came to my attention, and I promptly left the theater to
work out some of the details of my new plan. The next day I consulted Bush.
[56g, p. 112]
Scientists get their new ideas at odd moments, much as people in other
walks of life. In this respect Wiener was no different. The Englishman
Charles Babbage, one of the great minds in the history of computing, hit
upon the idea of automating computation also in curious circumstances,
as the following anecdote (c. 1825) reveals:
One evening 1 was sitting in the rooms of the Analytical Society at Cambridge, my head leaning forward on the table in a kind of dreamy mood, with
a Table of logarithms lying open before me. Another member, coming into
the room, and seeing me half asleep, called oW, "Well, Babbage, what are
you dreaming about?" to which 1 replied, "I am thinking that all these Tables
(pointing to the logarithms) might be calculated by machinery." {G 10, p. II}
Wiener, like Leibniz, liked to see his ideas worked out in practice. But myopic and manually clumsy, he had to rely on colleagues for
this. Dr. Bush put a student to tryout Wiener's idea "in-the-metal", and
in 1929 came a paper, "A new machine for integrating functional
products", by K.E. Gould, which opens with the words: "This paper is
intended to describe a new machine conceived by Dr. Wiener, which
performs certain integrations which are important in engineering, .... "
{GI2, p. 305}. Nearly a dozen papers and theses were written on his
machine by Bush's younger colleagues, starting in 1929 and ending with
Hazen and Brown {H5} in 1940, by which time the apparatus was
equipped with a moving film and called the cinema integraph.
The Wiener integraph is optical in nature. Radiation from a
plane source S of uniform intensity is made to pass through the aperture
in a plane opaque screen A, parallel to S, and very close to S and of the
same width a and the same height h, the aperture having the shape of
the region bounded by the (horizontal) x-axis, (the vertical) ordinates
x = 0 and x = a, and the curve y = f(x). This filtered radiation then
passes through the aperture in another plane opaque screen B, parallel
to A, and distance t away, of width al2 and height h, this aperture
having the shape of the region bounded by the (horizontal) x-axis, the
(vertical) ordinates x = 0 and x = a12, and the curve y = g(2x). But
this screen B is swung through 180 about the y-axis, and then slid
parallel to the x-axis a distance t12. The radiation from the second
aperture, so displaced, falls on a narrow vertical receiver R distance t
away from B, of height 3h, and very small width. The intensity of
illumination of R is nearly S:f(x)g(t - x)dx, the accuracy improving as
hit and alt decrease. For a diagram and proof, see Note 1, p. 175.
The Gould apparatus of 1929 {G 12} was superceded in 1931 by
another of T. S. Gray {G 14}, embodying several interesting technical
0
163
could be conveniently evaluated, by preparing a large number of cinema-frames for the functionsf( . , y 1), ... ,f( . , y n), attaching them on
film and stepping-in this film to the screen A. The whole operation and
recording was automated and fairly rapid.
These ingenious analog devices appear pathetically crude when
set beside our present-day analog machines. But they represented the
pioneering effort of the 1920s and 1930s to which the computer age at
least partially owes its existence. Actually, the integraph was put to all
sorts of uses during its lifetime. These ranged from the evaluation of
Fourier, Legendre, and other orthogonal series representations of functions to the solution of integral equations (Fredholm and Volterra)
appearing in different branches of engineering.
, real,
(1)
DO,
0], cf.
W(A)X(A) ,
A complex,
(2)
164
l3
and W(), called the frequency response function of the filter, has a
holomorphic extension! to the lower half plane. When the filter in
question is an electrical network with a finite number of lumped passive
elements, i.e., resistors, capacitances (storage tanks) and inductances
(coils), the input x(t) is an impressed voltage-difference, and the output
y(t) is a current, it turns out that
Y(Je) = R(iJe)x(Je),
i. e. W(),) = R (iJe),
(3)
1LJt)
I
e- t
k=o
0,
k)!
tk,
for
t ?:
0 (4)
for t < O.
165
=; [X} p~ ~~u)
du,
166
13
Vannevar Bush
1890-1974
ph yeA)
[co
10g~ ~~u) I
duo
f:()0
I log I
Y(A)112dA
f:oc
I ph
yeA) I 2dA.
167
Y.W. Lee
168
13
On his way to China with his family, Wiener spent two weeks in
Japan, and renewed contacts with S. Ikehara, his former student and
collaborator on the Prime Number Theorem, who had become a professor at Osaka University. Wiener also delivered lectures at Osaka and
Tokyo. The Wieners then headed for China. The first days were spent
in settling down, getting their young children enrolled in an Englishspeaking elementary school, and getting reacquainted with Lee and his
Canadian wife. Wiener and Lee then began work. In Wiener's words:
What Lee and I had really tried to do was to follow in the footsteps of Bush
in making an analogy-computing machine, but to gear it to the high speed
of electrical circuits instead of to the much lower one of mechanical shafts
and integrators. The principle was sound enough, and in fact has been
followed out by other people later. What was lacking in our work was a
thorough understanding of the problems of designing an apparatus in which
169
170
13
In 1940, with war raging in Europe, Dr. Vannevar Bush sought from the
MIT faculty research proposals that might contribute to national defense, as well as suggestions for its organization. Wiener responded
promptly. On the organizational question, he favored scientific collaboration which would cross scientific frontiers and would at the same
time be voluntary, thereby preserving a large measure of the scientists'
initiative and individual responsability. He accordingly suggested
... the organization of small mobile teams of scientists from different fields,
which would make joint attacks on their problems. When they had accomplished something, I planned that they should pass their work over to a
development group and go on in a body to the next problem on the basis of
the scientific experience and the experience in collaboration which they had
already acquired. [56g, pp. 231,232]
But bureaucracies, prone to big size, big budgets and static locations
tend to shirk off compactness, mobility and efficiency, and nothing came
of Wiener's suggestion. Moreover, as Wiener has noted [56g, p. 232],
171
many an inventor falls in love with his gadget, viewing it as his private
possession rather than as a small step in an ongoing evolution that
began with the caveman.
On the specific topic of war research, Wiener proposed the
design of an electronic computer to solve partial differential equations.
This, like the integraph, was to be a special purpose computer, designed
to carry out a specific mathematical task, viz. solve a PDE. But unlike
the integraph, it was to be a digital computer, which counts digits as we
do in our mental arithmetic, and does not measure quantities as in an
analog machine.
The earliest example of a digital computer is the abacus, which
until recently was used extensively in the Orient. The first mechanical
computers built by Pascal (c. 1642), Leibniz (c. 1672) and Babbage
(c. 1830), for addition, multiplication, etc. were digital, as were the desk
computers in vogue some decades ago. Digital computers are much
more accurate than the analog, and can yield as accurate an answer as
we need, if we are patient. But before World War II, they were notably
slower.
What prompted Wiener to make this defense proposal was the
question that Bush had raised with him in the early 1930s of the
possibility of mechanizing the solution of PDE's which, as we saw in
9A, are fundamental to field physics, and thus to a vast body of
engineering, and which are accurately solvable by pure mathematics
only in the rarest cases. Their mechanical solution to high degrees of
approximation was thus a vital practical need, and one of defense
potential in view of their appearance in naval and aeronautical engineermg.
In the PDE, the numerical data is spread over a planer or higher
dimensional region. In the Dirichlet problem, for instance, the data
consists of the values T(x, y, z) of the temperature at each point (x, y, z)
inside a room, cf. 7E. Wiener saw at once the futility of analog devices
in such a situation, and realized that it was crucial to replace the
function of many variables by one with just one variable. His 1923 work
with Phillips on the Dirichlet problem [23b] and his fine technological
intuition gave him a clue as to how this could be done in two steps. For
simplicity suppose that we have just two variables, i.e. the data are spread
over a planer region, say a square. First the continuous data must be
"discretized", i. e. replaced by the array of their values at a large number
of points that form a grid. Second, these discrete data must be read off
by a line-to-line scanning from left to right as in television (or indeed as
we do at a much slower pace when reading a page with our eyes).
If there are t points on each line of an n-dimensional cubical
grid, there will be en points in the grid, each of which will have to be
172
13
173
ber 21, 1940, and is reprinted in the Call. Works, IV, as [85a] and [85b]
respectively.2 Its abstract reads:
The projected machine will solve boundary value problems in the field of
partial differential equations. In particular, it will determine the equipotentiallines and lines of flow about an airfoil section given by determining about
200 points on its profile, to an accuracy of one part in a thousand, in from
three to four hours. It will also solve three-dimensional potential problems,
problems from the theory of elasticity, etc. It is not confined to linear
problems, and may be used in direct attacks on hydrodynamics. It will also
solve the problem of determining the natural modes of vibration of a linear
system. [85b, p. 125]
As D.K. Ferry and R.E. Saeks point out in the Call. Works, IV,
what Wiener proposed was a high-speed electronic special purpose
computer, modern in all respects except for a stored program. As they
say, Wiener's proposed machine employs (1) a discrete quantized numerical algorithm for solution of the PDE, (2) a classical Turing machine architecture, (3) binary arithmetic and data storage, (4) an electronic arithmetic logic unit, (5) a multitrack magnetic tape.
Wiener's appreciation of the novelty of the computational solvability of partial as opposed to ordinary differential equations appears in
the last section:
The fundamental difficulty in the solution of partial differential equations by
mechanical methods lies in the fact that they presuppose a method of
representing functions of two or more variables. Here television technique
has shown the proper way: scanning. or the approximate mapping of such
functions as functions of a single variable, the time. This technique depends
on very rapid methods of recording, operating on, and reading quantities or
numbers. [85b, p. 133] (emphasis added)
More specifically,
With magnetic scanning and printing, it does not seem too much to hope that
an entire set of reading, adding, and printing operations may be completed
in 10 -4 seconds. [85b, p. l30].
2 The documents together with comments by Drs. B. Randell, R. Saeks, D.K.
Ferry and the writer are reprinted in the Annals of the History of Computing
9 (1987),183-197.
174
13
As Ferry and Saeks remark, such speeds were attained only by the first
generation of transistorized computers (IBM 7090, 7094, etc.) in the late
1950s.
The report then alludes to the adaptability of the type of computer envisioned in domains in which the solution of important equations are so complex that they become almost impossible to solve
rigorously, and which "suffer greatly from the lack of computational
tools"-domains such as the theories of turbulence, shock waves, supersonic propulsion, explosions and ballistics [85b, p. 134]. In Section 2
Wiener illustrated in detail the scanning procedure for the parabolic
equation
(6)
175
What was extremely unfortunate was that Wiener's memorandum was not brought to the attention of the computer builders at
Harvard, Pennsylvania and elsewhere. (Dr. H. H. Goldstine, at least,
was unaware of the memorandum until the writer brought it to his
notice in early 1982.) In his autobiography Wiener has suggested that his
computer proposal was rejected because Bush's "estimation of any work
which did not reach the level of actual construction was extremely low"
[56g, p. 239]. But this explanation is not convincing, for, as we have seen,
Dr. Bush had treated Wiener's earlier tentative suggestions with considerable respect.
A more convincing explanation is that Bush felt that the line of
approach that Wiener had advocated, while drawing away manpower
from other important projects, would only marginally assist the war
effort. One has also to take into account the recent investigations of
Dr. Brian Randell {R I} of pre-war memoranda in the Bush archives in
Washington, which reveal that Bush was himself involved in an allpurpose digital computer project at that time-a fact barely known to
his colleagues. This relevation casts doubts on Dr. H. H. Goldstine's
suggestion that it was perhaps an over-commitment to the analog point
of view and a failure of vision that stopped Bush and his MIT colleagues
from pursuing electronic digital devices, until J. W. Forrester broke the
trend, cf. {GI0, p. 91}. The memorandum shows that in engineering
vision Wiener had hardly a peer, and could glimpse far into the future.
Note 1: The Wiener integraph (cf. 13A)
y'
R
Fig. 1.
176
13
narrow vertical strip, with center x/2 and height h, in which the wedge
cuts the screen B. But the part of this strip above height g(2 . x/2), i. e.
g(x), is opaque; hence only g(x)f(x)Ax of the radiation in the wedge gets
through the aperture in B and falls on the receiver R. Now rotate the
wedge about the midsection of R from the position x = 0 at A to
x = a at A. Then both apertures are fully swept in the rotation, and we
see that the total intensity at R is J~f(x)g(x)dx.
Next, imagine a contraption whereby the screen B can be made
to swing through 180 0 about the y-axis, and then slid parallel to the x-axis
a distance t/2 thus affecting the transformation xJ2 ---+ - x/2 + t/2. If the
illumination of the receiver R is measured with the screen B in this
position, the reading will clearly be the desired J:f(x)g(l - x)dx.
Note 2: Proof of the relation 13B(3)
c
Fig. 2.
LRC circuit
Each of the passive elements causes the impressed voltage x(t) to drop.
For simplicity, suppose that we have only one resistor, inductor and
capacitor of strengths R, L, C. (See Fig. 2). The voltage drops, according
to the Laws of Ohm, Faraday and others, are
Ry(t),
L/(t),
-1
Jt y(s)ds,
0
Jt y(s)ds
Ry(t)
+ -1
Ry'(t)
+ -1 y(t)
C ()
= x(t),
or equivalently
Ly"(t)
= x'(t).
177
13 Notes
whence
y(A)
iA
- - - - - - - - X(A),
L . (iA)2 + R (iJ.) + ~
C
),/(0."
R},
lie).
(ii) The Fourier transforms in are rational functions of the simple form,
cf. 13B, (5),
178
13
in(w)
Vi
(11 -+ iW)n
iw
1,
iw
Wreal.
Fig. 3.
An LR network
The voltage ratio for this, viz. R/(R + iLw), can be converted to
l/(i + iw) by suitable choice of the parameters Rand L. The other factor
(of absolute value 1, since w is real) is realizable by cascading n identical
"lattice" circuits of the type shown in Fig. 4.
L
~)
L
Fig. 4.
179
13 Notes
input
output
k~O
ckLk
i.e.
4u(x, y, t+ 1) = u(x+ 1, y, t) + u(x-l, y, t)
+ u(x,y+l, t) - u(x,y-l, t)}.
(8)
180
13
The scanning procedure that Wiener had in mind to solve (8) does not
lend itself to easy description. For simplicity suppose that the curve over
which the boundary conditions are prescribed is the periphery of a
rectangle in the xy - plane with sides parallel to the coordinate axes, and
that the lattice has v points in each row. Assume that these points are
counted from left to right, row after row, from top to bottom, and put
into a single sequence Pm n = 1,2, ... , v 2 . Then (6) yields for points
P n' not on the periphery,
u(P m t
1
1) = 4[u(P nH t)
u(P n -
1,
t)
u(P n -
v,
t)
This is because the points in the sequence just above and just below
P n are clearly P n - v and P n + v'
Thus if we have a record of the values of u(P n' t) on a linear tape, if we can
scan these values so as n to obtain u(P n + 1, t), u(P nT " t), if we have a rapid
mechanism for adding these values together and di~iding by four, and if we
can imprint this on the tape for the next run, leaving all boundary values
unaltered, say by some switch-off mechanisms, we can solve the system of
equations (4) (i.e. our (8)) mechanically. [85b, p. 129]
It is apparent from this that what Wiener meant by "scanning procedure" is a partial difference reduction adapted to transcription on a
tape. In [23b] (with H. B. Phillips) he had employed this very net to
attack the Dirichlet problem by a corresponding difference equation.
The idea goes back to Runge (1911) and reappears in the late 1930s in
the relaxation technique of D. G. Christopherson and R. V. Southwell,
cf. B. Randell Call. Works IV, pp. 135-136.
181
14
1940-1945
Soon after the shelving of his computer project, Wiener was drawn into
the study of the control of servo-mechanisms that was being carried out
by engineers at MIT. His work during World War II grew from a
suggestion he made to the servo-engineers in early November 1940, that
networks with frequency responses of a certain kind, into which the
positional data of an airplane's flight trajectory is fed, might provide a
means for evaluating its future locations, and so assist in the improvement of anti-aircraft fire. Anti-aircraft fire control was a problem of
tremendous military importance at that time because of German air
superiority over England.
The problem of anti-aircraft fire control deviates from similar
problems in land and naval ballistics in that the speed of the target
(airplane) is a substantial fraction of that of the missile fired at it.
Consequently, to score a hit the gun must be aimed at a future position
of the plane. The future path of the plane has therefore to be extrapolated from its observed path in the past before the firing can begin.
The extrapolation must be done by a high-speed computer, and the
aiming and triggering done automatically. Briefly, the gunner and his
range computation tables have to be supplanted by a speedier and more
accurate automaton that does its own radar tracking, anticipating,
aiming and firing.
The servo-engineers under the leadership of Professor
S. H. Caldwell went to work on Wiener's idea using the Bush differential
analyser at MIT. These efforts were voluntary, for the group had no
financial support from the National Defense Research Committee
(NDRC). But as Dr. Caldwell wrote in his proposal to Section D-2 of
the NDRC on November 22, 1940:
The preliminary results obtained ... indicate a much higher probability of
success, and it is the conclusion of all the members of that group ... that a
more extensive exploratory program should be initiated. [la, pp. I, 2]
182
14
183
Julian Bigelow
Photograph taken
in 1942
Very candidly, he advised that "all efforts in this field should be concentrated on features leading to more rapid production and simplicity
of use in the field" [IX, p. 8].
To backtrack a little, in January 1941, an engineer from IBM,
Mr. Julian Bigelow, was assigned to assist Wiener in his work. As
Wiener wrote in his Final Report to Dr. Weaver, Bigelow rendered
considerable help in the theoretical work as well as in numerical computation, "but his assistance was greatest and indeed indispensable in
the design of apparatus, and in particular in the development of the
technique of electric circuits of long time constant". (For more on
Bigelow's contribution and on the Lee-Wiener network, see Note 1,
184
14
p. 195.) In 1946 Bigelow became the chief engineer for the computer at
the Institute for Advanced Study, Princeton, that von Neumann was
building.
B Idealization of the flight trajectory and resulting tasks
185
f(t - r)dW,,(r),
and the (unknown) weighting W h was expressed by a simple expression involving the covariance function qJ of the signal! (c. January,
February 1941)
(b) They designed an electric filter that could realize this weighting W h
in-the-metal, i.e. which, when fed with an input voltage!, would
give out the voltage g", to a high degree of approximation.
(c. March-lune 1941)
(c) They built a contrivance which would generate "an entire family of
curves having the same sort of randomness as a maneuvering
airplane". (c. July-December 1941)
(d) They studied the statistics of actual airplane tracks, and for this
visited several army and navy centers in eastern United States where
such work was being done, receiving full cooperation from the
authorities. (The so-called "runs 303 and 304" made at the Antiaircraft Board at Camp Davis, in particular, assumed a conspicuous
role in the later phases of their work.)
C The mathematical theory
Let f(t) be the incoming signal from the airplane at the instant t. By
Wiener's ideal assumption f is in the class S, and with the ergodicity
186
14
+ 12(t),
II(t)
lit)
rpl(t)
rp(t)
rp22(t) - rpl2(t) ,
is also known.
Given h > 0, the problem is now to determine from the known
covariances rp, rpl' the weighting W h such that the average
Ih(t)
I(t - r)dWh(r),
(1)
T~oc
IT
-T
III (t
h) -
CXJ
h)
I~ rp(t - r)dWh(r),
min.
(2)
must satisfy
t;:::: 0.
(3)
II
12 on ( - oo,t)
implies
JI
J2 on ( - oo,t).
(4)
187
See Note 2, p. 195, for more on the equation (3) and its role in the theory
of instrumentation.
D Operation of the anti-aircraft predictor
188
14
********
Wiener saw to it that "the spot had to be moved by a control which was
complicated to begin with, and furthermore felt completely wrong" [56g,
p.250].
From operating with the random curve generator, Wiener and
Bigelow found to their surprise that it was "highly sensitive to individual
nervous reactions of the operator". "The recorded performance of each
operator showed an autocorrelation curve definitely similar to that of
his other performances, and differing from that of other operators"
[VII, p. 7, para. 2]. Wiener and Bigelow were thus led to believe that
although the pilot is ostensibly a free agent, the environment in which
he is placed during a bombing run effectively constrains his behavior.
His individual movements are auto- and cross-correlated, as a result of
which the course his airplane executes in the sky is itself autocorrelated.
To sum up, they realized that
... the "randomness" or irregularity of an airplane's path is introduced by
the pilot; that in attempting to force his dynamic craft to execute a useful
189
All this reinforced their initial supposition that the observed trajectory
of the airplane belongs to the class S that Wiener had demarcated in the
late 1920s.
190
14
Did the termination of the project lead to resentment on Wiener's part and to his estrangement with his engineering and military
colleagues? The documentary evidence is crystal clear.
As our quotations from his reports show, Wiener openly analyzed the limitations in his approach. Obviously he was the one best
qualified to know whether his project merited continuation. And on this
point he was quite clear, as the last page, marked "Recommendations",
from his Final Report of December 1942 shows:
The author finds that an optimum mean square prediction method based on
a 10 second past and with a lead of 20 seconds does not give substantial
improvement over a memory-point method, nor over existing practice. He
proposes to check this result with a similar method in which the more
unpredictable parts of the course are discarded\ and to discover the maximum lead for which his method yields an effective improvement in prediction. He also wishes to examine the theoretical effect of better tracking. He
considers that if these investigations do not yield a much more favorable
result than those already carried out, and he does not anticipate that they will,
he will have established that new developments in the design of long-time
predictors have already reached the point of diminishing returns, and that all
efforts in this field should be concentrated on features leading to more rapid
production and simplicity of use in the field. [IX, p. 8J (emphasis added)
In his follow-up letter of January 15, 1943, after describing the further
work that would be needed to improve flight prediction, and the merits
of perusing it, Wiener concluded:
I hesitate to make any such recommendation because I frankly do not know
to what extent work of this sort may tie up present working sources of the
country, and because the present expectation of great improvement is too
distant to be significant in the present war. I most definitely do recommend
that such study be made within some part of the long time program of our
armed services. [X, p. 5J (emphasis added)
Thus Wiener advocated ending attempts to design better flight predictors as a part of the war effort, but maintained that further studies of the
many interesting issues that had come up during the project would be
in the long-range interests of the armed forces.
Did the Bell engineers and defense officials who watched this
project feel that the effort was wasted? Far from it. As late as July 1,
1942, Dr. Stibitz, the Chairman of Section D2, noted in his diary: "I feel
more strongly than ever that any attempt to predict curved flight, which
is anything more than a stop-gap, must make use of Wiener's theory"
[VIII]. That this whole undertaking was favorably evaluated is clear
from the official Preamble, entitled Statistical method of prediction infire
control, that accompanied the limited circulation of Wiener's Final
This, by the way, was Stibitz's suggestion of February 12, 1942.
191
Report. Pointing out the limited value of straight line prediction (were
it to succeed, enemy aircraft would not fly straight courses), the Preamble continues:
Thus it is important to investigate how one can, by various methods, predict
other than a pure linear signal. For analytical curves of varying amounts of
complexity (horizontal circles, helices with dive or climb, etc.) the prediction
procedures can be worked out by simple geometrical methods. But to investigate prediction problems for signals which contain noise due to tracking or
to flight errors, or for signals corresponding to various general types of
curved flight, that is a much deeper and more difficult matter.
This is, in very general terms, the problem with which Professor
Wiener has been concerned. In a previous memoir (Report to Services No. 19)
he applied powerful analytical tools to develop a statistical method of predicting. In the present report he indicates, in two papers dated December 1, 1942
and January 15, 1943, the result of applying his method to certain definite
cases. That these particular applications did not turn out to be ofpractical importance does not, in our judgment, mean that the study was not well worthwhile. The
general theory will doubtless have other applications; and it was a matter of importance to know just how successful this statistical method would befor the antiaircraft problem. [Preamble to IX & Xl (emphasis added)
Professor Moore adds "I was a member of the System Analysis Branch.
Much of the work was classified, so I cannot cite appropriate references". But this was not all. It soon transpired that the nervous systems
of animals, big and small, are replete with these quickenings (Wiener's
"lead mechanisms" of 1940 or "anticipationary feedbacks" [6lc, p. 112])
as well as "lag mechanisms" to slow down, of which Wiener was also
quite aware.
G Secrecy, overwork and tension
Wiener was very appreciative of his exchanges with the engineers of the
Bell Telephone Laboratories from whom he benefited considerably, as
we just saw. He was much irked by security regulations which forbade
192
14
Actually, in this instance, as we gather from the same letter, the clamp
was put on Wiener not from the fear that he would spill some valuable
beans, but from the fear that his "highly mathematical" discussions
"might well have prevented the close get-together of the military and
civilian personnel which was a primary objective of this conference"!
But Wiener had a strong and lasting aversion to secrecy. Reflecting, later
on, on the overall meaning of secrecy in public affairs, he was to write:
This demand for secrecy is scarcely more than the wish of a sick civilization
not to learn of the progress of its own disease. [50j, p. 127]
193
194
14
Academician
Andre Kolmogorov
1903-1987
Notes
195
came not from any definite program on my part nor, I believe, from any on
theirs but was due to the fact that we had come into greatest activity at about
the same time with about the same intellectual equipment. [56g, p. 145]
In terms of our circuit diagram of the Lee-Wiener filter shown in Ch. 13,
Note 3, Wiener was referring to the improved designing of the instruments AI, ... , An"
Note 2: The equation 14C(3); theory of instrumentation (cf. 14C)
196
197
15
198
15
on my career. I like to think that, as Dr. Cannon was my scientific grandfather, so Dr. Rosenblueth is my scientific father. {G3, p. 24}
199
Arturo Rosenblueth
1900-1970
Photograph taken
in 1945
Courtesy of Mrs.
Virginia Rosenblueth
obvious when Dr. Rosenblueth's book Mind and Brain {R2} is compared
with more stereotyped tracts on the philosophy of science.
Dr. Rosenblueth's book clearly reveals that Wiener's philosophical mentors were also his own, and that a central concept in his scientific
philosophy was what Whitehead and Russell have called relation-structure. Affirming the position taken by Russell in his book Human Knowledge, Its Scope and Limits {R7} that all we can know of the objects and
events in the universe is their structure, Rosenblueth wrote:
I fully concur with this statement, and believe that this limitation is basically
due to the fact that as soon as the messages we receive from the outside reach
the afferent sensory fibers, they are phrased in a code which has nothing in
common with the original objects or events except for a common structure.
When a person hears a symphony, the messages sent by the orchestra reach
the listener as air vibrations. These vibrations stimulate mechanically the
receptors of the organ of Corti, and these receptors set up nerve impulses
along thc fibers of the VIIIth nerve. It is clear that at this stage the physical
events that are taking place are of an entirely different kind from those
200
15
Rosenblueth then demarcates the concept of structure in terms of isomorphism, precisely as done by the great mathematician H. Weyl {W5,
p. 25}. Briefly, a system comprising a set X and n relations R], ... , Rm
thereon (each Rk being an r-tuple of elements of X, r :2: 1 depending on
k)] has the same structure as another system comprising a set Yand n
relations S], ... , S n thereon, if and only there is a one-one function T
on X onto Y such that
(x], ... , x r ) e Rk
if & only if
Thus the relational-predicates "has the same structure as" and "is
isomorphic to" are synonymous.
By way of illustration, Rosenblueth cites the large number of
such isomorphic (or similarly structured) systems involved in the transition from the set of events that constituted the composition of the
piano sonata, Opus 111, by Beethoven in 1823 to the set of events in the
mind of a listener today of a Schnabel recording of the sonata made in
1932. Here the relational structures that remain invariant are the ratios
among the fundamental frequencies of the simultaneous or successive
sounds, their intensity, duration, timing and the like. And an important
transformation T is between these and the musical notation on sheets of
paper (coding), another important T being the final mental decoding
that enables us to appreciate the sonata. But for optical perception,
among the relations that must be preserved are also some that are
topological in nature. In all such transitions from sets of events in the
outside world to those in our minds" ... the only features that we can
perceive are those which remain invariant under the physiological transformations which our organisms impose on them" {R2, p. 57}.
This strong endorsement of some of his own early philosophical
ideas by an active and independent neurologist had a tonic effect on
Wiener. From Rosenblueth's seminar Wiener became conscious of the
importance of brain wave encephalography as a tool for understanding
the mind. Also, Rosenblueth's thought on the preservation of structure
gave a foundation to the theory of pattern-recognition that McCulloch,
1 We may look upon properties as unary relations.
201
Pitts and Wiener advanced around 1945 ( 16C). But the first input from
Rosenblueth came in the course of Wiener's war work in the stormy year
1942 (next section).
202
15
203
Wiener was speaking of his first visit to Mexico in 1945 2 at the invitations from the Mexican Mathematical Society to speak at their meeting
in June of that year, from Dr. Rosenblueth of the National Institute of
Cardiology to continue their collaboration, and from his former colleague Dr. Vallarta, then in the Science Commission (Comisi6n
Instigadora y Coordinadora de la Investigaci6n Cientifica). Wiener
spent ten weeks in Mexico, during which he collaborated with Rosenblueth in an area in physiology proper. Wiener's biological curiosities,
held dormant since 1912 when he shifted from the study of zoology to
that of philosophy, now found expression. He could now show what his
"teachers did not recognize", that despite all his "grievous faults", he
might still have a contribution to make to biology. (See the full quotation on p. 40-.) With encouragement from Dr. Walter B. Cannon, they
sought to find out how muscular contractions during epileptic convulsions
are related to movements of the heart [46b]. (To appreciate the physiolog2 The "Mexico, 1944" in the title on p. 276 of [56g] should read "Mexico,
1945".
204
205
cal responses, i.e. the slow spreading of the local effects, elicited by
stimuli, along seemingly random paths without any noticeable periodicity, cf. {R2, pp. 40-42}. It was their hope that some of their theorems
could be subjected to experimental tests, and the range of theoretical
conjecture thereby reduced.
Although well aware that this 60-page work lacked adequate
computational analysis, the paper was presented at the mathematical
meeting in Gadalajara, and in the following spring (1946) at the first of
a series of semi-annual meetings organized by Dr. Warren McCulloch
with the support of the Josiah Macy Foundation. These meetings,
lim ted to a group of about 20 scientists, were for the presentation of
fresh research, and informal discussion. The paper which embodied this
work [46b], was to serve as a catalyst in the work of Drs. W. J. Osher and
T. Cairns in the mid-1960s, and its influence continues even to this day.
For details see the comments of Garcia Ramos in Call. Works, IV.
D Muscle clonus
In the summer of 1946 Wiener again went to Mexico, this time with
support from the Rockefeller Foundation. He teamed up with Rosenblueth and his colleague, Dr. Garcia Ramos, to analyze experiments on
the extensor thigh muscles of a cat in order to study clonus: "the familiar
spasmodic vibration which many people experience when they sit crosslegged with one knee under the other" [56g, p. 277]. This phenomenon
has to do with feedback. Rosenblueth felt that their results were substantial enough to be presented to colleagues, and they were read at the
Macy meeting in New York in the fall of 1946 but not published. Their
experimental confirmation "did not come up to Arturo's rigid requirements" as Wiener has explained, cf. [56g, p. 278]. Recently, Dr. Garcia
Ramos, the only surviving author, has found the supporting evidence,
and his updated version [85c] appears in the Call. Works, IV, along with
his commentary.
The least technical account of their experiments, which were on
the exterior thigh muscles of decerebrated cats under anesthesia, is in
Wiener's Cybernetics:
We cut the attachment of the muscle, fixed it to a lever under known tension,
and recorded its contractions isometrically or isotonically.4 We also used an
oscillograph to record the simultaneous electrical changes in the muscle
itself ... The muscle was loaded to the point where a tap would set it into a
periodic pattern of contraction, which is called clonus in the language of the
4
206
15
Wiener's words, "We cut ... " must be taken with a grain of salt. As
Dr. Garcia R-.amos explained to the writer, Wiener would come to the
laboratory and watch them do their experiments, throwing in an idea or
two. Then, following further talk on the results, Wiener would retire to
work out a mathematical explanation.
The experiments showed that the greater the mass of the moving
system, the greater the tension required to initiate clonus. On the other
hand, the frequency of clonic oscillations is nearly independent of the
mass. Continued observations, however, revealed slow changes in the
basal tension and amplitude. The many other experimental results are
too technical to describe.
Theoretically, Wiener and his colleagues approached the neurophysiological circuit involved in clonus from the standpoint of communications engineering. Almost half of their 50-page manuscript is
devoted to the application of the mathematical theory of communication to the problem. A preliminary question was the determination of
the physiological concepts that should correspond to the different ones
such as current or voltage in engineering. And since the circuit in
question is highly non-linear, a major concern was the determination of
a suitable linearization.
The fact that the rhythmic activity maintains itself, after the
stimuli that induce it have ceased, shows that a closed feedback circuit
is involved in clonus. The authors assume that it is made up of one
afferent nerve and one efferent nerve, besides the muscle and kinesthetic
receptors. If the number of impulses transmitted per second by the
efferent nerve is taken for the "current", then the circuit is highly
non-linear. But it becomes nearly linear if the logarithm of this number
is taken for the "current". To quote from their manuscript:
In the clonus studied, the envelope of the motor impulses recorded electrically in the eat's quadriceps, and inferentially in the efferent nerve, has the
form:
207
quantities in which the oscillations could possibly be linear. This is equivalent to considering the nervous processes to be not additive but multiplicative. [SSe, p. 501]
With the logarithmic basis, the oscillation frequencies deduced from the
data with the aid of the principles of servo-mechanisms turned out to be
good approximations to the observed clonic frequencies.
Encouraged by this experimental confirmation, the authors advanced a physiological explanation of clonus in their paper. We may
sum it up by using Wiener's words:
There is a strong suggestion that though the timing of the main arc in clonus
proves it to be a two-neuron arc, the amplification of impulses in this arc is
variable in one and perhaps in more points, and that some part of this
amplification may be affected by slow multineuron processes which run
much higher in the central nervous system than the spinal chain primarily
responsible for the timing of clonus. This variable amplification may be
affected by the general level of central activity, by the use of strychnine or of
anesthetics, by decerebration, and by many other causes. [6lc, p. 21]
Wiener also felt that the observed slow change in basal tension and
amplitude could be studied by the method of secular perturbations in
celestial mechanics due to H. Poincare and G. W. Hill. The final part of
the chapter on feedback in Wiener's Cybernetics [61c] is devoted to this
question.
E The spike potential ofaxons
To regularize their collaboration, Wiener and Rosenblueth induced
MIT and the Mexican Cardiological Institute to approach the Rockefeller Foundation for support of an arrangement whereby Wiener would
spend one semester every other year in Mexico City, and Rosenblueth
a part of the intervening year at MIT. This was readily obtained, thanks
largely to Dr. Warren Weaver, and Wiener spent the fall of 1947, 1949
and 1951 in Mexico, and Rosenblueth made several visits to MIT.
A project that Wiener and Rosenblueth undertook in Mexico in
1947 was to determine the shape of the fluctuation, y = f(t), of the action
(or "spike") potential at a fixed point x on a stimulated axon, to find out
if the shape depended qualitatively on the axon, and how it is influenced
by experimental factors such as temperature. This investigation was
done in collaboration with Dr. J. Garcia Ramos and Walter Pitts [48c].
It involved considerable experimentation, followed by mathematical
description based on the empirical material as well as on theoretical
premises.
Wiener and his colleagues removed intradural portions of the
dorsal roots, 2 to 4 ems. long, from the lumbo-sacral segments of the
208
15
spinal cord of cats,S and extracted fine strands containing several nerve
fibers. They applied rectangular electric pulses to these strands, and
obtained electro grams of the response, picked up by electrodes that fed
via an amplifier to a cathode ray oscilloscope. Appealing to the all-ornone law (Note 1, p. 214), they concluded that whenever the latency of
response varied but the shape of the curve f remained unchanged, only
a single fiber was active.
After rectification of various errors, they were led to the hypothesis that the stimulus-response relationship is "piece-wise linear", i.e.
that the time-domain can be broken into three consecutive intervals in
which the stimulus obeys different linear differential equations. This
suggested that different factors were at work during the three intervals.
The team then tried to account for the observed results by introducing
dynamical partial differential equations connecting the current along the
axon with the potentials inside and outside each cross-section. (For
more details, see Note 2 on p. 215.)
This work was superseded four years later by the more revealing
(non-linear) partial differential equation for the ionic current due to
A.L. Hodgkin and A.F. Huxley, but its pioneering aspect is worth
stressing. It was the first to bring in field conceptions, continuous time
and PDE's, in axon theory-an area where the overall, all-or-none
behavior, is clearly discrete. Even today the subject lacks powerful
general principles governing the physiochemical mechanism, and the
modus operandi, consisting of educated guesswork based on careful
experimentation, is not very different from that followed by Wiener and
his co-workers in 1948. Some of their conclusions received experimental
confirmation in the work of Dr. 1. Tasaki in 1956.
F The statistics of synaptic excitation
A fourth project that Rosenblueth and Wiener undertook in Mexico was
on the statistical analysis of the input-output relation in spinal reflexes
in dogs and cats [49b]. This too was done in collaboration with Dr. Garcia Ramos and Walter Pitts, and it involved a good deal of experimentation as well as statistical hypothesizing and testing.
In a reflex action the nerve impulse engendered by the stimulation of a receptor organ travels along a spinal nerve to the spinal cord,
5 The lumbar-sacral segment of the spinal cord is the segment that extends
from the middle to the lower back, and is thus farthest from the brain. The
dorsal root of a spinal nerve (which in quadrupeds is located on the upper
side of the cord) is composed of incoming sensory neurons. The intradural
portion is the part inside the Dura Mater, i.e. the outermost of the three
sheaths (the meninges) that surround the cord and encase the cerebrospinal
fluid.
209
210
15
211
vous channels, e.g. hormone transmission by blood circulation. Associated with such feedback are the homeostatic diseases, like leukemia,
where often
... what is a fault is not so much an absence of all internal control over the
process of corpu~cle formation and corpuscle destruction but a control
working at a false level. [56g, p. 292]
212
15
Notes
213
8 Kinesthetic organs (also called proprioceptors) are receptors within the body
which gather infornlation about the internal environment, much as the sense
organs gather information about the external environment.
214
15
A neuron will fire when and only when the magnitude of its
stimulation exceeds the threshold. When it fires, the resulting
nerve impulse is independent of the excess. Thus a neuron has
only two states:firing or quiet. (This is referred to as the "all-ornone" law.)
(2) The speed of a nerve impulse for a given axon is fixed, being
dependent on its cross section.
(3) A nerve impulse along an axon is either excitatory or inhibitory,
depending on whether it tends to encourage or discourage the
firing of the neuron or muscle fiber on which its terminus impinges.
(4) The chemicals produced at the terminus of an axon by the
propagation of a nerve impulse take a certain time (the synaptic
delay) to diffuse across the synapse and reach an adjacent neuron.
(5) The trans-synaptic excitation of a neuron by such diffusion
requires an accumulation over a time period (temporal summation), or the simultaneous arrival of the material diffused from
more than one axon (spatial summation).
(6) Time is required for an excited neuron to return to normal, i.e.
to its resting PD between the center and periphery of crosssections of its axons. During this refractory period, the neuron
will not fire, even if subjected to above-threshold stimulation.
Notes
215
Nerves are cords (or bundles) made up of many nerve fibers, i.e.
neurons with their dendrites and axons, often as many as 100. Their
length can vary from a fraction of an inch to several feet, and their
thickness from that of a thread to that of a pencil. The nerves have a
distinguished end, called the center. Afferent nerves are ones in which the
nerve impulse proceeds towards the center. Efferent nerves are those in
which nerve impulse propagates away from the center. The nerve impulse is carried in each of the separate fibers in the nerve. Mixed nerves
are those having fibers made up of sensory neurons, as well as fibers
made up of motor neurons. Such nerves can transmit impulses in both
directions without interference, much as traffic on a two-way street.
The portion of the eNS necessary for carrying out the involutary
processes vital to survival (e.g. the beating of the heart) is called the
autonomic nervous system. It comprises two complete and separate
systems of nerves, called the sympathetic and parasympathetic nervous
systems. A nerve from each system runs to each body organ, and their
impulses have a balancing effect on the organ. The sympathetic nerve
speeds up the organ in its activity, the parasympathetic nerve slows it
down. The advantages are those of a double feedback.
Note 2: Spike potential ofaxons
Let y = fit) be the shape of the fluctuation. Then, after correcting for
various artifacts, Wiener and his co-workers found that the curve /
"rises, inflects, reaches a maximum, and then declines with only one
more inflection" [48c, p. 279] (see Fig.). They also found that the second
differences that approximate the second derivative /n were "quite
irregular" [48c, p. 283]. It was reasonable to assume that the transformation leading from the stimulus to the response / was only piecewise
linear, i.e. the domain of/was made up of intervals over which/ obeyed
different linear differential equations, the points t of division being those
at which / and f' are continuous but!"
is not.
,
216
15
Be-PI
t < to
to ::;; t ::;; t [,
t[ < t.
y = Ae~ "t,
y = I - Nt
y = Ce- yt
< 0
o::;;t::;;'Jr
r > t,
t
Be - pt
(1)
I - B,
Arx
B/3 - N,
Cye yr
N - Be- Pr .
(2)
217
Notes
that the firing threshold is attained at the inflectional point to, and the
refractory period commences at t 1, (cf. Note 1).
To account for these results, Wiener and his co-workers introduced a dynamics, involving the potentials v Jx,t) and vo(x,t) inside
and outside the x-cross-section of the axon at instant t, as well as the
longitudinal currentj (in the direction of increasing x) governed by the
differential equations [48c, p. 293]
aj
at
-A(v - v.)
0' '
av
at
-Rj,
av.
"
at,_- Z'j,
218
16
219
220
16
The analytic engine weaves algebraic patterns just as the Jacquard 100m
weaves flowers and leaves. {GIO, p. 22}
Now many different fabrics can be had with the same pattern by altering, for instance, the colors of the warp threads. Likewise, different
numbers can be computed from the given algebraic expression by assigning different values to the constants a, b, etc. and the variables x, y,
etc. Thus to accomplish a specific algebraic task, the control unit has to
prescribe a sequence comprising both operations, i.e. steps of the algorithm, as well as the assignment of values for the constants and variables.
The Babbage engine, were it constructable, would indeed be a
machina ratiocinatrix as conceived by Leibniz. This is because any
mathematical or logical problem, capable of mechanical solution, can be
resolved into a repetition of fundamental steps, and the exact routine to
be followed laid out in the control unit as a sequence of instructions.
With remarkable insight Leibniz and Babbage visualized this possibility,
even though they had no way to render scientific the intuitive idea of
mechanical solvability. This became possible only in the 1930s after the
rise of mathematical logic, and the explication, by K. Godel, A. Church
and S. C. Kleene, of the idea of recursion latent in the Principia M athematica of Whitehead and Russell (cf. SB). But this idea had to be
rendered in a form serviceable to the design of computers. This step, of
going "from the calculus to a system of ideal reasoning machines"
(Wiener's words) was taken concurrently by E. L. Post and A. M. Turing
in 1936.
Computability, Turing's explication of the pre-scientific notion
of mechanical solvability is a bit too complex to explain. We shall
instead describe a hybrid Turing machine which should give the reader
an inkling of Turing's idea. The machine comprises a moveable tape
divided into cells, in any of which a single symbol from a fixed finite
alphabet can be inserted. Its engine is capable of carrying out just a few
simple operations 0), O2 , 0 3, . , such as reading the symbol in the fixed
front location, or replacing it by another symbol, or moving the tape one
step to the right or to the left. The algorithm to be used to solve a given
mathematical or logical problem is prescribed by a program, i.e. a
sequence of instructions I), 1 2, 1 3, . In to be carried out in some order
on the machine at the times t = 1,2,3, ... Each Ii is of the type:
(1) Perform operation Ok> and then proceed to instruction Ii' (the k and
j depending on i),
(2) perform operation Ok and depending on whether the symbol in the
front cell is a or h or ... , turn to the instruction IiJ or If, or ... ,
(3) stop.
For instance 17 might be: "Move the tape a step to the left and then turn
to instruction 14". (For details, see Note, p. 238.)
221
Success with computers has engendered the faith that given any
solvable mathematical or logical problem, a program can be written
whereby it can be solved on the Turing machine, i.e. that this machine
accomplishes the tenable part of the calculus ratiocinator of Leibniz.
The next problem, difficult in its own right, is to engineer the
operations involved in the Turing algorithm effectively, and get the
machina ratiocinatrix. The key step in this, as far as electronic machines
are concerned, was C. Shannon's discovery in 1938 that the operations
of the Boolean algebra of propositions such as alternation, conjunction
and denial (in symbols v, 1\, ~) are realizable by electrical switching
networks {S6}.2 Let "true" symbolized by 'I' be represented by an
electric pulse, and "false" symbolized by '0' be represented by its absence. Imagine that a pulse, applied to one of the switches S), S 2, . . in
a circuit, closes it. Then the truth table of v, is realizable by a simple
network, with switches "in parallel" as the following diagram shows:
v
0T01
1
I 1
The output indicated by the needle will be I only when a closing pulse
i), i2 is applied to one or both of the switches S), S2, exactly as in the
truth table for v. Likewise, the truth table for 1\ is realizable by a
simple network of switches "in series" as the next diagram shows:
1\
0T00
1
I0
Notice that in the binary scale, the multiplication and addition tables are
x
~
I
+ 10
I0 I
I0 1
and
~
1 11 10
The first matches the table for 1\, and the second differs from the table
for v only in having 10 in place of 1 in the last place. But the symbol
2
222
16
'10' can easily be had from the symbol '1' by the Turing operations of
moving the tape one unit to the left, and then imprinting '0' in the front
cell. In the electronic computer more advanced techniques involving
semi-conductors are used, but as the reader may guess the logic in the
procedure is that of switching. Thus the Turing machine, so equipped,
is capable of mechanically performing the fundamental operations of
both logic and arithmetic, and is the all-purpose machina ratiocinatrix
of Leibniz.
With this understanding of the nature of computers, Wiener was
able to augment his early ideas on the nature of logic. Important logical
concepts such as proof, contradiction, and undecidability are expressible
in machine language. More and more, he came to regard logic as the
idealized theory of logic machines. Like our minds, these machines
suffer from all the "non-removable limitations and imperfections" affecting all natural mechanisms and organisms [6Ic, p. 125]. "All logic is
Warren S. McCulloch
1897-1969
223
224
16
Walter Pitts
1923-1969
225
B,
For the visual modality, Ais a layer of the visual cortex, and the group
G is homomorphic to the group of Euclidean motions in physical space.
McCulloch and Pitts suggested that the nervous system is
equipped with neural nets to affect the formation of the transforms T(A)
of an apparition A. Let
(1)
Then Ais duplicated on N -1 cortical sheets AI, ... , ANI. The neuron
x in Ahas N-l axons leading to the neurons Tj(x), ... , T N-l(X) on the
sheets .;1(1' ... , AN-I, respectively. The apparition A, i.e. the state of
excitation of the cortical manifold JIt, will then yield in succession the
transformed apparitions Tl(A), ... , T N-l(A) having the same Gestalt as
226
16
u {S \ Tk(A)}
falls below a certain tolerance. How apparitions are stored in the memory is not yet understood, nor in fact is the mechanism of long-time
memory.
Another type of net for pattern recognition is a teleological
mechanism operating on the principle of negative feedback. For any
apparition A, let A among its replicas T(A) be deemed standard. Letf =
if), .. ,fr) be a multi-parameter which determines the state, i.e,f), ... ,
fr be real-valued functions on f'l' such that if for each k, fk(A) = fk(B),
then A = B. Let I . I be a norm on IR" (cf. 7C). The mechanism
computes the deviation
0 0 = If(A) - f(A) I,
IflT(A)} - 1(A)
< 00 ,
227
This suspicion was to affect much of Wiener's work on electroencephalography, as we shall see ( 16F below).
228
16
229
ries that in human speech there are only about five important frequency
bands. To use the words of the authors:
In this method a sound is carried from a microphone to a bank of filters. The
output of each of these filters, except possibly the low frequency stage, is
rectified and used to modulate a low frequency carrier and the output of
these carriers as well as of the possibly by-passed low frequency stage, is
carried to a number of stimulators, which may be properly designed electrodes or may be electromagnetic vibrators resting on suitable points of the
skin, such as the five fingers. [4ge, p. 512]
230
16
signals generated by
the biceps and triceps
muscles in the upper
arm.
Both photographs
courtesy of Mr.
T. Walley Williams,
Liberty Mutual Insurance Company,
Boston, MA
Doctors and engineers unveiled today what they called the "Boston arm", an
electronically operated artificial limb that an amputee flexes simply by willing
it to flex, as is done with a natural arm.
In part, the Boston arm project grew out of talks in 1960 between
Dr. Glimcher and the late Dr. Norbert Wiener of M.I.T.
A good deal of the work on the arm was actually done under the
direction of Professor Robert W. Mann and his group at MIT. The
"fingers" of this arm are activated by electric stimuli (EMGs) received
from the elbow muscles (biceps and triceps), and as the newspaper
pointed out, the arm was "volitional": its use did not demand any special
initiative by the user beyond the normal appetency.
Wiener was neither a professional physiologist nor a surgeon nor
an engineer, and his specific recommendations on prosthesis sometimes
proved wrong. Professor Mann has remarked (Coil. Works, IV) how
Wiener underestimated the potentialities of the tactile modality for
prosthesis. The best modern aids for the blind rely on this modality
rather than on hearing. Also Wiener's idea to communicate with the
deaf by skin stimulation channeled from a few audio-frequency bands
was oversimplified and did not work out. Some of this specific suggestions on limbs for the amputee were also off-par. His belief that implant-
231
232
16
short papers-the philosophy of science, science, mathematics, brain research, and engineering, all converging on applied prostheses design! {Coil.
Works, IV, p. 439)
To Wiener, the term prosthesis had a very wide connotation. The human
species suffers from severe neurophysiological and other limitations. A
device, such as the telescope, that enables humans to overcome a limitation, is prosthetic in Wiener's wide sense. As the quotation above
indicates, Wiener visualized a prosthetic role for the electronic computer
evcn in 1940, i.c. before it came into existence. This vague thought
received a much sharper enunciation in 1946 when John von Neumann
became profoundly interested in the computer from the broadest standpoint. Thus J. von Neumann and H.H. Goldstine wrote:
... really efficient high-speed computing devices may, in the field of nonlinear partial differential equations as well as in many other fields which are
now difficult or entirely denied of access, provide us with those heuristic hints
which are needed in all parts of mathematics for genuine progress. In the
specific case of fluid dynamics these hints have not been forthcoming for the
last two generations from the pure intuition of mathematicians, although a
great deal of first-class mathematical effort has been expended in attempts to
break the deadlock in that field. To the extent to which such hints arose at
all (and that was much less than one might desire), they originated in a type
of physical experimentation which is really computing. We can now make
computing so much more efficient, fast and flexible that it should be possible
to use the new computers to supply the needed heuristic hints. This should
ultimately lead to important analytical advances. {Gil, pp. 4-5}
233
Wiener had been interested in brain rhythms ever since he had learned
of them in the 1930s from Dr. Arturo Rosenblueth. They had been
discovered by the British physician R. Caton in 1875, and recorded by
the German physiologist Hans Berger in 1928. When a pair of electrodes
are placed at two points on the scalp, a faint and fluctuating potential
difference (PD) can be detected. This detection is impossible with ordinary instruments, but with the combined use of a powerful amplifier and
a cathode ray oscilloscope, a record of this (amplified) PD as a function
of time, a so-called electroencephalogram (EEG), can be obtained.
When the advent of amplification brought these EEGs into
vogue, there was hope that they would tell a useful story about the brain.
But the curves turned out to be extremely irregular ("a drunkard's idea
of a roller coaster" [57e, p. 115]), and even the most skillful and experienced readers were able to decipher only a few crude rhythms. But
Wiener, like some others, was certain that the EEGs were indeed "writings of the brain", worthy of serious study.
Wiener tried to convey the importance of the EEGs to an audience of medical men in New York by asking them what would happen
if two electrodes were placed on the wall of the lecture hall:
234
16
I would find small differences of potential ... The actual size of the potential
wouldn't mean much to me. I don't know what the walls are made of, I don't
know what electric wires are behind them ... , but I can tell you something
that I do pick up ... Since this is New York City which has sixty cycle
alternating current, I should pick up sixty cycle hum ... This frequency of
sixty times a second would be very reliable, as reliable as the regulation of
the current of the voltage in the city power stations, and this is quite
reliable .... Similarly, when we are examining potentials in the brain ... we
cannot expect that the actual size of the observed potentials is very significant. We can expect, however, to get very accurate results from thc timing
of them. [57e, p. 112]
Actually the MIT auto correia tor was based on better mathematical
principles than Grey Waiter's machine. It embodied Wiener's ideas from
1930: it was in principle a Michelson interferometer ( 9F). The EEG f
was put on magnetic tape, and played back on an apparatus having two
playback heads, thereby creating the output fit) + fit + h), the delay
235
W. Grey Walter
191O~1977
JTJ(t + h )J(t)dt,
u
236
16
... there is a clock mechanism here that can be definitely picked up from the
brain; it beats ten times per second and it takes 50 seconds to lose or gain a
beat. [57e, p. 116]
This "clock" is affected by the individual's constitution and physiological condition, for the alpha rhythm
disappears in deep sleep, and seems to be obscured and overlaid with other
rhythms, precisely as we might expect, when we are actually looking at
something and the sweep rhythm is acting as something like a carrier for
other rhythms and activities [61c, p. 142].
237
speeds up slow alternators and slows down fast ones, and brings into
play extra alternators when the load is excessive. Wiener worked out a
mathematical theory to account for this sort of phenomenon in [58i]
which is too difficult for this book, and is described in {M9,
pp. 120-121}.
Wiener attached much importance to this idea of non-linear
entrainment of frequency. On it was based his understanding of selforganizing systems, especially those where a rhythm is organized, as is
the case with nearly all living systems. The cells of such a system are both
senders and receivers of messages. The cells are non-linearly coupled, so
that the frequencies of time-variation of important parameters governing the cells can pull one another together (as in the brain) or push one
another apart, thereby creating in either case an organization.
Such a system as it gathers greater and greater synchronism will emit an
impulse which has a greater and greater tendency to synchronize oscillators
which have not already been pulled into place, until by a mass action they
constitute a definite pulsating organ. [58f, p. 13]
238
16
the toughest. McCulloch admitted that his own theories of the neurophysiological mechanism were intended largely to suggest fruitful experimentation which hopefully would reduce the need for wide-ranging
theoretical conjecture in this difficult subject. Wiener tended to view his
theories on electroencephalography in less conjectural terms. If, however, we gauge his contribution in more modest terms, there is little
doubt as to its penetrating and wholesome quality.
Note: Structure of the Turing machine
01:
02:
0 3:
04:
by 1.
239
17
Wiener's intellectual collaboration with Rosenblueth, Bigelow, McCulloch, Pitts and earlier with Lee, Bush and Russell, and his work on
electronic computing, predicting and filtering embraced several fields of
science. But through this work ran a unifying thread centered on the
concepts of message, noise, feedback, communication, teleology and
control.
The concept of message was not restricted to messages of human
origin-those involved in the flashing of fireflies and those flowing
through electrical machinery were also included. The question of intelligence, which led back to logic, and to the possibility, discerned by
Leibniz and Pascal, of logical automata and machine intelligence were
thus included.
The same sort of scientific experience was shared by a group of
scientists, many of whom had been in touch with Wiener. Thanks to the
generous support of the Josiah Macy Foundation, this group was able
to get together from time to time in different places. In 1943 it decided
to name itself the "Teleological Society". Led by J. von Neumann and
Wiener, the society included eminent engineers, computer theorists and
physiologists, among them Rosenblueth, McCulloch, Pitts and Bigelow
of whom we have already spoken, and Dr. H. H. Aiken of Harvard, Dr.
H. H. Goldstine of the University of Pennsylvania, builders of the Harvard Mark I and ENIAC, and Dr. Lorente de No of the Rockefeller
Institute, among others. It gradually dawned on the members of the
group that the problems which concerned them, in which message and
the other above-mentioned concepts are cardinal, constituted a new
interdisciplinary field of scientific activity.
The "cybernetical circle", as the group came later to be known,
had made substantial contributions to science during the 1940s. They
had analyzed the notion of noise, demarcated the concept of information and demonstrated its negentropic thermodynamic significance,
designed filters that could intelligently predict the future values of
stationary time-series, discovered the servo-mechanical aspects of vol-
240
17
untary purposeful activity, developed a calculus for the neuronal process and a theory of Gestalt or pattern-recognition, and recognized the
common mode of operation of the brain and the computer.
But by the late 1940s there was a growing feeling among many
that the movement was losing steam and had not addressed itself to a
truly major problem then on the frontiers of science. As Dr. W. Grey
Walter was to say later in 1969:
So often has a cybernetical analysis merely confirmed or described a familiar
phenomenon in biology or engineering, so rarely has a cybernetical theorem
predicted a novel effect or explained a mysterious one. {WI, p. 94}
241
242
These two great minds thus had much in common, despite very
obvious differences in their thought processes and presentations: von
Neumann's being thoroughgoing and immaculate to the last detail,
while Wiener's were diffuse and sometimes almost incomprehensible,
and despite their widely differing temperaments and administrative
propensities. The impression that an intrinsic barrier separated their
mathematical attitudes, which a reader of Professor Heims's recent account {H6} may perhaps get, would be quite mistaken. True, von Neumann's great mind had an axiomatic bent, whereas Wiener's relished a
"more tighter attitude" (cf. 7C). True that one could cope with several
different problems during a stretch of time and that the other could not.
Also true that the words "crazy", "cranky", which might prop up
half-affectionately in talking about Wiener, were never applied to von
Neumann. But these differences have small significance. Their important
ideas overlapped, even to the extent of sensing the prosthetic use of
computers. Thus it was natural, if not inevitable, that sooner or later
there would be a healthy interaction between the two.
Wiener's active interest in von Neumann's work began with the
ergodic theorems in 1931, cf. 9F, 12B, and in the early 1940s von
Neumann was struck by Wiener's recognition of the proximity of the
brain and the computer, and his teleological viewpoint.
C Von Neumann's letter on the direction of cybernetical research;
molecular biology
243
244
17
245
246
cal substitution techniques to vary the x-ray pattern. I realize that this is in itself a big
order, and that it is still by a factor 10 3 off our goal-but it would probably be more
than half the difficulty.
In addition there is no telling what really advanced electron-microscopic
techniques will do. In fact, I suspect that the main possibilities may well lie in that
direction. The best (magnetic) electron-microscope resolutions at present are a little
better than 10 A = 1 mil. With 4 x 10 6 atoms in a volume of3.7 x 10- 17 cm 3 = 3.7
x 10 4 mill, the average atomic volume is 10- 2 mill, and hence the average atomic
distance about 1/5 mil. Hence the 1 mil resolution is inadequate-but not very far
from what might be adequate. A resolution that is improved by a factor of 10-20
might do. It is dubious whether electron lenses can be improved to this extent. On the
other hand, the proton microscope need not be more than 2-4 years in the future, and
it would certainly overcome these difficulties.
Besides all these developments might be pushed and accelerated. Of course,
everybody knows what a 1-1/2 A resolution would mean: One could "look" at an H
atom, and with a little more, say 1/5 A, one could "see" the Schroedinger-chargecloud-of the orbital electrons. But the physiological implications are even more
extraordinary, and they should receive a great deal of emphasis in the immediate
future.
At any rate, I think that we could do these things:
Study the main types of evidence: Physiology of viruses and bacteriophages, and all
that is known about the gene-enzyme relationship. (Genes are probably much like
viruses and phages, except that all the evidence concerning them is indirect, and that
we can neither isolate them nor multiply them at will.)
Try to learn a reasonable amount about the present state of knowledge and
opinions concerning protein structure.
Study the methods of organic-chemical structure determination by x-ray
analysis and Fourier-analysis with their necessary complement of manipulations.
Study the principles and methods of electron-microscopy, both in the direction of electron optics and in the direction of object-manipulation. Try to get oriented
to the possibilities of proton-microscopy.
Finally: Compile for our common use two lists: (1) Relevant publications,
with the main emphasis for the immediate future, considering our lack of education,
on books and survey articles. (2) Persons from whom we might learn most about the
state of affairs and the outlook in these fields.I did think a good deal about self-reproductive mechanisms. I can formulate
the problem rigourously, in which Turing did it for his mechanisms. I can show that
they exist in this system of concepts. I think that I understand some of the main
principles that are involved. I want to fill in the detals and to write up these considerations in the course of the next two months. I hope to learn various things in the course
of this literary exercise, in particular the number of components required for selfreproduction. My (rather uninformed) guess is in the high ten thousands or in the
hundred thousands, but this is most unsafe. Besides, I am thinking in terms of
components based on several rather arbitrary choices. At any rate, it will be necessary
to produce a complete write-up before much discussing is possible.
Certain traits of the gene-enzyme relationship, of the behavior of some
mutants, as well as some other phenomena, seem to emphasize some variants of
self-reproductivity, which one would be led to investigate on purely combinatorial
grounds as well. E.g.: Self-reproductivity may be symbolized by the schema A-A.
What about schemata like A .... B .... C -+A, or A .... B -+ C .... C, or A .... B .... C
.... D -+ E .... C, etc.?
247
1903-1957
This is as far as my ideas go at this moment. I hope you will not misinterpret
the anti-neurological tirade at the beginning of this letter. Of course I am greatly
interested in that approach and I have the greatest respect for the important results
that have been obtained in that field, and in our border area with it. I certainly hope
these efforts will continue. I wanted to point out, however, that I felt that the decisive
"break" was more likely to come in another theater. I was trying to formulate and to
systematize my motives for believing this, and the simplest literary mode to do this
is the controversial one. I hope therefore, that I have not given you a false impression
of the spirit in which I am starting a "controversy".
I am most anxious to have your reaction to these suggestions. I feel an
intense need that we discuss the subject extensively with each other.
Hoping that this letter has not been unbearable just by its sheer length, and
hoping to hear from you and to see you again soon, I am, with the best regards,
Yours, as ever,
248
17
249
250
17
251
18
Cybernetics
252
18
Cybernetics
steersman, and began using the term to refer loosely to the theory of
communication and control in the animal and the machine. Wiener was
surprised to learn later that over a 100 years earlier the French physicist
Andre-Marie Ampere, an early leader in electromagnetism (after whom
the unit of electric current the "amp" is named) had coined the term
"cybernetique" to refer to the "art of government" {A3}, (1838, 1843),
and that in the same year, 1843, the word "cypernetyki" appeared in a
Polish work on management by S. Trentowski in Poznan.
Ampere's cybernetique and Wiener's cybernetics are far apart,
however, as Ampere's Chart on the classification of the sciences and the
place allotted to cybernetics testifies.
cosmological
sciences
~
/noological proper (including philosophy, psychology,
noological
pedagogy, in the science on literature, etc.)
~
ethnological (also including history)
social
physico-social (social economy, mipolitical
litary art)
"'1
~
253
tion is central: the impact of sea waves on the ship's hull, applause and
jeers from the audience. Was then Plato a precursor of cybernetics? The
answer is unclear, for we know of no place where he affirms that the
back-and-forth flow of messages is what unites navigation and rhetoric.
We may tend to read too much in his use of the term, forgetting that it
had a long career:
The word "XVpepWXW" and its derivatives are fairly frequent in Homeric
poems as well as in Ancient literary and philosophic Greek writings. The
primary meaning of this family of words denotes the steering of a vessel, or
guiding the vessel in a specified direction (Odyssey, 3, 283) and also driving
chariots (Plato, Theages, 123 c). Several writers use this verb with a metaphorical tint to denote "guiding" or "governing". {G4, pp. l44-l45}
In this century, a clear awareness of the centrality of "reversibility", i.e. what we all call feedback, is shown in the book Psychologie
Consonantiste (1938, 1939) written by Stefan Odobleja, a military physician in Romania, {O1}. He attempted to build the science of psychology
around the concept of consonance, i. e. a continuing confluence of factors
maintained by feedback. This psychological starting point, and Odobleja's incognizance of engineering feedback, resulted in his blurring of the
distinction between positive and negative feedback, and in his interpretation of feedback coupling (or reversibility) as being energetic rather
than informational. Notwithstanding these limitations borne of time
and place, Odobleja asserted the important role of consonance and
reversibility in phenomena outside psychology, for instance in biological
processes, bringing a fresh light to bear on the work of French physiologist Claude Bernard (1879). Where the informational aspect in a feedback is negligible, his descriptions are weighty. For instance, his theory
of attention, i.e. of bringing and holding an object in mental focus, is
rather like Wiener's treatment of visual feedback (cf. 16C).
Odobleja's role in the history of cybernetics is well described by
the title of a recent collection of essays: Odobleja Between Ampere and
Wiener {D7}. Ampere emphasized regulation or control, Odobleja the
feedback loop, and Wiener and Shannon the message, as well as the
connection between all three elements.
The cybernetical attitude carries with it much more than an
appreciation of control by informational feedback, however. Among its
other central ideas are:
(1)
(2)
(3)
a universality that assigns only a secondary status to the autonomy of the different sciences;
a recognition that intelligence extends into the inanimate world;
an indeterministic causality in which the universe is a cosmos,
but of an incomplete order that permits teleology and freedom;
254
(4)
18
Cybernetics
255
Henri Bergson
1859-1941
256
18
Cybernetics
Wiener's concept of transducer subsumes Ashby's definition of a machine, as a function on S x E to S, where S is the set of internal states
and E the set of external conditions {AS, p. 242}. For, such a machine is
merely a transducer for which "incoming messages" and "outgoing
messages" are members of S x E and S, respectively. Ashby's conception is very fruitful. As he explains:
A machine is that which behaves in a machine-like way, namely, that its
internal state, and the state of its surrounding, oefines uniquely the next state
it will go to {A5, p. 251}.
257
message
nOIse,
as well asfeedback. The laws governing the signal are more complicated
and harder to handle than those governing the message. The efficacious
way to solve a practical problem is to extract the message, i.e. to filter
out the noise, to deal exclusively with the uncontaminated message (even
though it occurs nowhere in nature in isolation), and finally correct the
error due to idealization by taking into account the presence of noise
and of transients. There are, of course, many classical problems in which
the noise is so small that it can be regarded as absent. For purposive
systems the analysis of the effect of the signal on the receptors of the
system and the internal feedback of the outputs within it are important.
Another task is the coding and decoding of messages so as to minimize
the influences of noise and maximize the flow of information.
Cybernetics thus brings to the forefront concepts and ideas
relevant to Bergsonian time such as noise, entropy, feedback and purpose, message and control, one or more of which have appeared in
specific fields such as thermodynamics or servomechanism engineering,
but which have been absent from the vocabulary in which the scientific
method is ordinarily presented. To this extent, cybernetics is the extension of the scientific methodology necessitated by the existence of processes for which time is Bergson ian.
C Is cybernetics a science?
In the preface to the 1961 edition to his book Cybernetics [48f], Wiener
speaks of cybernetics as no longer a program but an "existing science".
There is some uneasiness among scientists as to the nature and domain
of this science. (See, for instance, J. R. Pierce {P5, pp. 208-209}. Its
status and domain are discernible, however, from the discussion in the
previous section. Cybernetics is a branch of mathematics, with its own
specific primitive terms, machine, signal, noise, information, etc., much
like hydromechanics with its own specific terms, mass, pressure, vortic-
258
18
Cybernetics
Just as the objects of geometry are idealizations of real objects, and just
as a strange geometry with no immediate application to real objects may
merit serious consideration and prove later to be of great practical value,
so too with the study of mechanisms. This thought is again masterfully
conveyed by Dr. Ross Ashby:
Cybernetics, then, is indifferent to the criticism that some of the machines it
considers are not represented among the machines found among us. In this
it follows the path already followed with obvious success by mathematical
physics. This science has long given prominence to the study of systems that
are well known to be non-existent-springs without mass. particles that have
mass but no volume, gases that behave perfectly, and so on. To say that these
entities do not exist is true; but their non-existence does not mean that
mathematical physics is mere fantasy; nor does it make the physicist throw
away his treatise on the Theory of the Massless Spring, for this theory is
invaluable to him in his practical work. The fact is that the massless spring,
though it has no physical representation, has certain properties that make it
of the highest importance to him if he is to understand a system even as simple
as a watch. The biologist knows and uses the same principle when he gives
to Amphioxus, or to some extinct form, a detailed study quite out of
proportion to its present-day ecological or economic importance. {A6, pp. 2,
3} (emphasis added)
This trenchantly expressed view-that the key to practical efficacy lies in the study of the non-existent - is Pythagorean-Platonism
259
260
18
Cybernetics
Replace "complex control system" and "data" by Wiener's "multipleinput, multiple-output transducer" and "message" respectively, and we
are, in essence, back to the Wiener-Ashby definition. Furthermore,
Soviet cyberneticians have been able to incorporate a good deal of
Wienerian thought in their own framework. It is rather the overtones of
the word "data", the near-exclusive concern in Soviet cybernetical literature with digital data and algorithmic considerations, and the omission
of analogue devices that go against the attitude of universality (cf.
l8A (1 and are confining.
For Soviet cyberneticians the narrower interpretation of cybernetics offered certain advandages. Because of the officially imposed ban
on mathematical logic in the Soviet Union until the late 1950s, they had
a good deal of catching up to do in the subject, and so an extra emphasis
261
Less ontology-bonded Marxists did not fall into this trap. Wiener
informs us that in 1947 Professors J.D. Bernal, J.B.S. Haldane and
H. Levy, all distinguished British Marxists, "certainly regarded it
(cybernetics) as one of the most urgent problems on the agenda of
science and scientific philosophy" [61c, p. 23].
In the late 1950s the official Soviet attitude towards cybernetics
sobered considerably. Wiener was invited to the Soviet Union in 1960,
treated with respect, and was requested to address the philosophical
section of the Soviet Academy of Sciences, and expound his views in the
periodical Voprosy Filosofii, [61b]. Cybernetics came to be actively
pursued in the Soviet Union, but with the narrower conception we
mentioned. The narrowness runs the danger of fostering over-specialization, reducing communication between cyberneticians and workers in
cognate fields (excluded from cybernetics by definition), retarding the
application of cybernetics to industry, and inhibiting the synthesis of a
cosmical or even global cybernetical perspective, commensurate with the
needs of contemporary civilization.
262
18
Cybernetics
E Ontogenetic learning
Cybernetics, in the full sense of Wiener and Ashby, has wide scope, for
definable within its framework are conceptions of great consequence
such as purpose, learning and self-organizing. Of these the concept of
purpose, broached in 15B, is fundamental.
Wiener realized that once the concept of a purposive mechanism
is defined (cf. 15B), that of a learning mechanism can be explicated.
Every machine has an input-output transformation T defined by setting
T(j) equal to the output when the input is f In a purposive mechanism
there is, moreover, an ideal transformation To, which is cherished but
unrealized, i.e. T(j) only approximates To(j) to some degree. A purposive mechanism is said to learn, if it has the ability not only to perform
the transformation T, but also to change T from time to time so as to
bring it closer and closer to the ideal transformation To, i. e. to reduce
the difference I T(j) - T()(j) I . Wiener visualized such a learning mechanism as a system comprising a performing filter A coupled by feedback
to a non-linear filter B [61c, p. 173; 64e, pp. 14,20-21]. A carries out the
routine of transforming inputfinto output T(j). B keeps a record of past
inputs, outputs, and errors of performance of A, and has devices for the
re-evaluation of the parameters governing this performance. Periodically, the system takes "time-out" to make this re-evaluation. The results
are automatically fed back to A, and the routine is resumed with
improved efficiency. For instance, the component B of a learning antiaircraft battery would record the long-time trajectories of incoming
airplanes, and compute therefrom estimates of the covariances of the
hypothetical underlying time series, cf. 14C. For this, B would have to
include non-linear devices such as square law rectifiers. When the improved estimates of the covariances are fed into component A of the
battery, the new T(j)'s produced by A will be closer to the ideal predictor
To(j) than before, i.e. the performance will improve.
Wiener believed that although our understanding of learning
machines is in its infancy, chess playing automata embodying the principles just outlined would be realizable in the future [50j, p. 176-177].
These automata, of "higher type" (cf. p. 52), would be superior to the
existing machines based on Dr. Shannon's designs {S8}.
Wiener compared this ability of a mechanism to learn, i. e. to
improve its performance by appropriately modifying its response transformation T in the light of environmental realities, to ontogenetic learning among biological organisms. He believed that in the higher animals
the parameters which determine T are synaptic thresholds (which control the firing of neurons) and that the "re-evaluation" or "learning"
consists in changing these thresholds for the better. He also felt that in
263
264
18
Cybernetics
W)*f
W 2*(fxf)
W 3*(fxfxf)
+ . ..
(1)
J~ W)(t - r)f(r)dr
J~ [W 2(t -
gives its "quadratic part" and so on. Thus, the single convolution
weighting W of the linear network, cf. 9B (11), gives way to a sequence
ofweightings WI' W 2 , of higher and higher orders. Such expansions
had already been studied by Volterra in the 1900s. The problem now was
to synthesize the higher order convolutions W z*, W 3*, ... by extending
the type of circuitry used in the Lee-Wiener network for synthesizing the
simple convolution W*.
2
265
From the very outset, however, Wiener was dissatisfied with this
scheme, By a remarkable intuition, he replaced the f by a Brownian
movement input, and took the convolutions of WI, W 2 , etc. with the
chaos determined by the Brownian movement, 12E(6). Thus he replaced the last two integrals by the chaotic integrals
J~WI(t -
r)dx(r, IX),
J~ J~Wlt -
IXE[O, 1],
which he had first introduced in the early 1930s, x(,) being the
Brownian motion, cf. 7D(3). The response of the network to such a
Brownian input signal x( . , IX) is now itself the signal g( . , IX) of stochastic process. This type of study of the non-linear network is made in
Wiener's 1942 report, referred to above. Wiener wanted to take the shot
effect as the random input since it offers the best physical realization to
the ideal Brownian motion, cf. 7D.
What prompted Wiener to use shot noise as a probe for the filter?
It was his belief that to "know" a black box it suffices to observe its
response to the most chaotic signal conceivable, to wit, the Brownian
motion. Wiener seems to have had this insight at least by 1930, for in his
memoir [30a] we read:
Imagine a resonator-say a sea-shell-struck by a purely chaotic sequence
of acoustical impulses. It will yield a response which still has a statistical
element in it, but in which the selective properties of the resonator will have
accentuated certain frequencies at the expense of others. [30a, p. 215]
Thus the response to "purely chaotic" impulses will "reveal" the resonator or filter, i. e. it willl give us enough clues about its character, perhaps
enough to construct its duplicate. Indeed, this is the case with timeinvariant linear filters, for one very important result in [30a] asserts that
if
get, IX)
J~CXJ w(t -
r)dx(r, IX),
IXE[O, 1],
then for almost all a, the response signal g( . , IX) has the spectral density
I W( .) 12 by bombarding the
filter with shot noise, and passing its response through a spectrophotometer. For a causal filter ( 9B) we can then retrieve W from
I W() 1,3 and thence W.
For linear filters, however, it is not necessary to use such random
probes, for the character or sinusoidal inputs are just as revealing, as we
266
18
Cybernetics
emphasized in 9A. But for non-linear filters we are left with the
Brownian motion as the most effective probe. This thought, which
dictated Wiener's choice of a probe, is stated lucidly in Wiener's posthumously published book:
The output of a transducer excited by a given input message is a message that
depends at the same time on the input message and on the transducer itself.
Under the most usual circumstances, a transducer is a mode of transforming
messages, and our attention is drawn to the output message as a transformation of the input message. However, there are circumstances, and these
chiefly arise when the input message carries a minimum of information, when
we may conceive the information of the output message as arising chiefly from
the transducer itself. No input message may be conceived as containing less
information than the random flow of electrons constituting the shot effect.
Thus the output of a transducer stimulated by a random shot effect may be
conceived as a message embodying the action of the transducer.
As a matter of fact, it embodies the action of the transducer for any
possible input message. This is owing to the fact that over a finite time, there
is a finite (though small) possibility that the shot effect will simulate any
possible message within any given finite degree of accuracy ... That is, if we
know how a transducer will respond to a shot-effect, input, we know ipso
facto how it will respond to any input. [64e, p. 34] (emphasis added)
Thus Wiener felt that random probes have a revelatory role in electrical
engineering that is similar to that of probes such as ink-blot tests,
random interrogation, etc., in human psychiatry.
This radical viewpoint suggested to Wiener the abandonment of
the Volterra scheme (1), in favor of a much deeper one, latent in his own
work on the homogeneous chaos in the late 1930s. However, some
systematic work by R.H. Cameron and W.T. Martin {CI} and by
S. Kakutani {K3} had to occur before Wiener could spot the latency,
and take off. The basic ideas are those of harmonic analysis, but the term
"harmonic" has now a much wider connotation than that in 9A.
Wiener considered a black box or filter B with an input terminal
and an output terminal, the internal structure of which is unknown. We
assume, however, that if at instant t it receives the shot noise (or
Brownian motion) signal x(t, 0:) at its input end, then its response g(t, IX)
at the output end can be read. (Here 0 ::0;; IX ::0;; 1, cf. 70.) In short, we
know only the transfer operator T of the filter B. Wiener showed that
if the filter B is time-invariant and causal (in the sense of 9A, B) and
stable (in the reasonable sense explained in Note 2 on p.270), but
otherwise quite non-linear, then Tis expressible in one and only one way
as a senes:
(2)
267
the WI' are standard white operators, i. e. they are transfer operators of a
one-input, one-output filter, the circuitry of which is fully known and
standardized. The summation is over all the p. The equality in (2) is to be
taken in the sense of root-mean-square error. Roughly speaking, not for all
inputs f, but only for all typical ones, will T(f) be equal to La I' Wp(f).
Wiener showed that for any such black T, the coefficients a p can
be determined. Once the ap's are found, we can get the circuit for apWp
by merely juxtaposing multipliers (amplifying/attenuating devices) to
the circuit for WI" We then build the huge circuit A by hooking up all
the apW I' circuits, according to the recipe Lap WI" This A will be a white
offspring of the black box B. That is, A will have the same transfer
operator as B, and thus be identical to its parent in overt behavior, but
it will have a fully blueprinted circuitry that may be quite different from
the internal structure of B, concerning which, we know nothing.
All this must be taken with many grains of salt, for obviously not
all the infinitely many apWp can be hooked together. Not only must the
series in (2) be conveniently truncated, but the formidable formulae for
the a I' impose more approximations. Thus different offspring of the same
B will in practice show variations much as in a biological population.
This process, leading from B to A, can be represented by a block
diagram with a unit, corresponding to the analysis that yields the crucial
coefficients uP' and the other that corresponds to putting the pieces
together:
---1
3-
-,
Analyzer
(/ff-
Synthesizer
-A
If from the diagram we omit B and A, we are left with the diagram of
a genetic scheme or incubator for transducer reproduction.
To return to phylogenetic learning, we now consider an assembly
of teleological learning mechanisms SI' S2, .... Each of these automatons S; is capable of the ontogenetic learning considered in l8E. But
we now imagine that they are also offspring to a common ancestor, and
can themselves reproduce via the agency of the incubator. We then get
a growing geneology of automatons that itselflearns as time flows, i. e.
we have mechanically simulated phylogenetic or racial learning.
G The distinction between scientific inquiry, and stratagem or contest
Raffiniert ist der Herrgott. aber boshaft ist er nicht. (A. Einstein)
268
18
Cybernetics
ing its objective. But B is present; after A has made the move a), B makes
a counteracting move b). The second move a; that A will now make will
depend on b) and will differ from a 2 . Likewise the third move a~ that A
will make will depend on the second move b 2 of B, and so on. Thus by
virtue of B's presence, A is obliged to adopt a new sequence of moves
, , ,
a), a2, a 3 ,
We may describe such a sequence of moves of a purposive
mechanism, which depend on the countermoves of another such mechanism with a contravening goal, as afight or contest. A rather common
type of contest occurs when A's objective is to control the purposive
mechanism B itself, and one of B's objectives is not to be so controlled.
We then have a fight or contest between two adversary mechanisms.
The process of adopting increasingly efficient policies for winning a contest is called strategic learning. As to the concept of strategy
itself, it has been axiomatized in the important work of J. von Neumann
and O. Morgenstern {V7}. Wiener, guided by Russellian typology, classified strategic policy into tactics, stratagem, superstratagem, etc.,
cf. 20C, D.
A purposive mechanism A makes an inquiry of another mechanism B, purposive or not, when its objective is to gain information
concerning B. Such an inquiry may just be part of the larger objective
of doing research, i.e. of determining the laws governing B.
A most vital aspect of scientific inquiry and research is that,
barring a very few insignificant exceptions, it involves no contest with
the mechanism being studied. Nature is not an adversary in the relevation of its structure, and there is no room for stratagem in such a study.
This is the thrust of Einstein's aphorism "Subtle is the Lord God, but
capricious He is not". This sentiment goes back to Heraclitus, 500 B.c.:
The Lord, whose oracle is at Delphi, neither reveals nor conceals but gives
tokens.
Nature neither assists nor restricts investigation. The PythagoreanPlatonic tradition, to which cybernetics belongs, rests on this firm faith.
The few exceptions to this rule are well known. A person may
resist giving out data on his health. A secluded tribe may resist an
anthropological investigation. Laboratory animals may behave abnormally in captivity, and to this extent foil the scientists' objectives.
The contrast between inquiry and contest is reflected in our use
of language. We adopt one mode of discourse to facilitate communication, another to administer contests and conflicts. The types of confusion to which the two forms of discourse are subject are quite different
[50j, p. 193]. That affecting the first stems from human error and/or
Notes
269
Transducer-reproduction bears some interesting similarities with physical, biochemical and biological reproduction. The reproduction scheme
or incubator of 18F cannot be used, because the large modifications
required change it beyond recognition. This is inevitable since cellular
tissue has a fine structure that is attuned for reproduction to a much
higher degree than the metallic pieces of our incubator. Nevertheless,
certain similarities subsist.
In gene reproduction, for instance, nucleic acids determine the
laying down of a chain of amino acids in the form of a pair of helices.
When these helical pairs split, each gathers to itself the molecular
residues needed to form its partner, and we get two genes instead of one.
Here the substrate, the nutrient medium, which is capable of assuming
a large number of forms, is made to assume the particular form of a
gene, by the presence of one such gene. Thus, the indeterminate substrate and the single gene playa role that clearly resembles that of the
indeterminate white box A and the black box B of transducer theory,
respectively.
When it comes to biological reproduction, the deviations from
our reproduction-scheme are even more pronounced. Apart from sexuality, which is absent in transducer-theory, biological reproduction is
"Mendelian", i.e. acquired characteristics, such as responses learned
ontogenetically, are not genetically transmitted; but they are so transmitted in our "Lamarckian" transducer-reproduction. Furthermore,
social heredity, i.e. vital responses learned after birth by imitation of or
training from live parents, and memorized, which is so important to the
life of the higher animals, is absent in our transducer-reproduction
scheme.
270
18
Cybernetics
Here the Lk are Wiener's favorite Laguerre functions, cf. 13B(4), and
the Hpk are the well-known Hermite polynomials.
The Wp filter is indeed "white", i.e. it is synthesizable. For each
k from to m, a cascade oflattice circuits, as in the Lee-Wiener network,
will deliver L k *. To form the polynomials Hw all we need are square-law
rectifiers (which turn fit) into JUn, scale-amplifiers and summing circuits. To form their product we can again fall back on square law
rectifiers and adders, by dint of the equation
J(t) get)
1
2
+ g(tW
-[ {jet)
- J(t)2 - g(t)2].
Let the black operator Tbe time-invariant, causal and also stable
in the sense that for each t,
00.
(4)
Then the coefficient ap in its expansion (2) is given, for almost any a in
[0, 1], by the time-average:
ap
lim -1
A
A~80
So
-A
(5)
Multiplier &
integrator
Shot-noise
generator
Black box
Notes
271
272
19
Wiener's thinking on automatization and its socio-economic consequences was far ahead of its time. Guided by his clear perception of the
electronic computer (cf. his 1940 memorandum, l3D), and his sharp
understanding of feedback, derived from his war work and physiological interests ( 14E, l5B), he had concluded by about 1942 that a
high-speed computer, properly programmed, could be made to run an
automatic factory. In his words:
... the modern ultra-rapid computing machine was in principle an ideal central
nervous system to an apparatus for automatic control; and ... its input and
output need not be in the form of numbers or diagrams but might very well
be, respectively, the readings of artificial sense organs, such as photoelectric
cells or thermometers, and the performance of motors or solenoids. With the
aid of strain gauges or similar agencies to read the performance of these
motor organs and to report, to "feed back," to the central control system as
an artificial kinesthetic sense, we are already in a position to construct
artificial machines of almost any degree of elaborateness of performance.
[6Ie, pp. 26, 27] (emphasis added)
273
274
19
dated July 26, 1950. In this letter Wiener did not really comment on thc
Reuther proposal but expounded his own geopolitical viewpoint. What
effect, if any, this letter had on Reuther is not known. (It is reproduced
in full in 20B, where Wiener's geopolitical thought is discussed.)
The Wiener-Reuther correspondence resumed on February 5,
1952, when Reuther wrote:
Our Union would be honored to have you speak at the session of the
Conference during which we will discuss the subject "Achieving and Maintaining Full Employment in a Free Society".
Leon Keyserling, Chairman of the Council of Economic Advisors
will share the platform with you and will speak on likely economic developments in the future. [MC, 146]
But Wiener declined, stating that he had had a hectic time in France and
Mexico and was "under doctor's orders to take it easy and rest".
So ended, it would seem, the contacts between Wiener and
Walter Reuther
1907-1970
Courtesy of
Ms. Norma Bogovich.
UAW.
Photo by
Alex.ander Archer
(CAW Solidarity
Magazine)
275
Reuther. They had only a marginal impact on the UAW, and nearly
none on the labor movement.
Neither Wiener nor Reuther foresaw the moral disability latent
in a "trade-union consciousness" that operates without a "revolutionary
consciousness", to use Lenin's phraseology. It imposes a cringing shortsightedness on both laborers and their leaders. All aspects of the employer-employee relationship (even production of dubious utility) are
held sacrosanct, as long as the worker can enhance his short-term
economic gains. Lenin described its effect on the Russian worker around
1900 in the words: "A kopec added to a ruble is nearer and dearer to him
than any socialism or any politics" IL 7, p. 69). For the UA W workers
of the latc 1940s (who some years later were to flood the market, without
a qualm, with flashy tail-finned gas guzzlcrs), one could say in the same
vein: "A quarter added to a dollar was nearer and dearer to them than
an understanding of the dislocation that might hit them 15 or 20 years
later as a result of the ongoing automatization".
Apropos of Wiener's efforts to alert labor on the socio-economic
consequences of automatization, a letter he received from a migrant
laborer, J. Printise Womack, is of considerable interest. Gnlike the
situation we just described in which Wiener had to go after labor leaders,
here it is a laborer who on his own initiative asked Wiener to intervene.
The letter indicates on awareness of the limitations of isolated tradeunion consciousness and of the impingement of cybernetics on religion,
at a time (1950) when Wiener had hardly begun to think in these terms.
Here is the letter [Me, 122]:
15 Aug. 1950
Dr. Wiener,
In the introduction to Cybernetics you indicated that you made an effort to
reach the labor unions and interest them in the problems which cybernetics poses.
As a farmer child, young man farm worker in this highly industrialized
farming area (the San Joaquin Valley), I wish it were possible for you to be heard by
the President's Commission on Migratory Labor, which is now holding open hearings
in various parts of the country.
It seems rather obvious that the industrial revolution on the farm is making
most farm workers and many farms obsolete. And the shadow of cybernetics, etc'.
hangs over them.
I believe the particular situation or the migrant in our success-ethic society
affords us a unique opportunity to get a hold on our selves. As a migrant worker or
producer he's on his way out. As a Migrant, as an individual, an indefinable, an
aloneness, he is valuable, if we choose to value ourselves as such.
Of course, from migrant to \1igrant requires a transcending motion-requires us to go beyond our identifications with religion, nationality, politics, race, etc.
to an open attitude toward man.
In any case, in view of cybernetics it would seem futile to try to integrate the
displaced farm workers and farmers into our success ethic way of life.
276
19
Too the promises of Stalinism to the peasants and workeJ;s"of the world are
apt to back fire on them. Industrialization will spread to the farms; cybernetics, etc.
to the factories, etc. Shall we continue to work along "integrating" lines, me-tooing
the Communists, or should we pose a vision of man as Migrant, or some such, hold
out the hope that individual Migrants of the world shall be relatively free to devote
themselves to their true work of finding out (trying to find out) who they are?
The address of the President's Commission on Migratory Labor is: 1400
Pennsylvania Avenue, N.W., Temps V, Washington 25, D.C. The Commission was
directed to inquire into social, economic, health and educational conditions, among
migratory workers, both alien and domestic, in the U. S. etc. Your work is relevant
to these problems. I hope you will be heard.
J. Printise Womack
277
278
19
of strange analogies, and of the distinction between inquiry and stratagem, cf. 18G. His unhistorical attitude debilitated his efficacy as a
reformer. Unappreciative of the important role of moral evil in shaping
history, he tended to assume uncritically the contemporary presupposition on the achievability of unlimited human progress by education.
Dewey's faith was inspiring. He wrote:
I see no ground for criticizing those who regard education religiously. There
have been many worse objects of faith and hope than the ideal possibilities
of the development of human nature, and morc harmful rites and cults than
those which constitute a school system. Only if all faith that outruns sight is
contemptible can education as an object of religious faith be contemned. This
particular form of faith tcstifies to a generous conception of human nature
and to a deep belief in the possibilities of human achievement, in spite of all
its past failures and errors. {D4, p. 148}
John Dewey
1859-1952
279
came in giving the devil his due, and on the nature of the actual outcome
ofaction and reaction. "The paradox of homeostasis is that it always breaks
down in the end" Wiener wrote, and T. S. Eliot has reminded us that "sin
grows with doing good" {E6, p. 44}. Dewey did not realize the havoc
that vested interests could cause by claiming his discipleship. The religion
he so eloquently described, brought on precisely what he did not wish:
... a mass of dogmas called pedagogy and a mass of ritualistic exercises
called school administration. (D4, p. 149}
Science, or the study of nature, must not be the center of education, for,
... apart from human activity, nature itself is not a unity; nature itself is a
number of diverse objects in space and time, ... {D4, p. 8}
-the study of science is educational in so far as it brings out the materials
and processes which make social life what it is. {D4, p. II}
Indeed, the product turned out by the public schools often failed to meet
minimal Deweyian or cybernetical standards. Those who crossed the
hurdle of the three R's, often tended to view science as an unedifying
game played against Nature in order to wrench out its secrets for
personal gain. Weak in geography and history, largely ignorant of
280
19
281
282
19
283
strophes have been great enough to wipe out every single commercial undertaking of an antiquity of thousands of years, and if they had not taken place,
the rate of interest for a long-time investment could scarcely have been twotenths of a per cent. [62c, p. 30] (emphasis added)
284
2.
3.
4.
19
285
To a scientist, and especially a humane one, who saw through the pitfalls
of both capitalism and socialism, it was an intellectual challenge to work
out a socially-responsible, free enterprise, devoid of the dislocative,
hedonistic and anti-intellectual propensities of capitalism, a system
whose strong homeostasis could absorb technological transitions.
Wiener's non-linear theory of self-organizing systems met with
difficulties in its applications to brain-wave encephalography, and unfortunately he made no attempt to apply it in the economic realm. To
go by his conversations, he had in mind an economic set-up vaguely
resembling the power station of an electric grid, in which "the total
generating system acts as if it possessed a virtual governor, more accurate than the governors of the individual generators ... " [61c, p. 201].
That is, the economy should consist of a number of semi-autonomous
units (some of which may be work-motivated rather than profitmotivated, or publicly owned) coupled to a central agency which continually corrects, or even intervenes more directly, to reduce the volatility of the market game, and so approximates a homeostasis conducive
to steady economic growth.
Wiener did engage himself a little more concretely on economic
questions while at the Indian Statistical Institute in Calcutta during
1955-1956. Social and economic planning were among the prime concerns of the Institute. Wiener felt that he could give a rationale to these
problems, and exchanged ideas with Professor Jan Tinbergen, the distinguished econometrician from Holland, who was also staying at the
Institute. It is best to let Wiener speak:
In any economic situation there are certain factors beyond our
control which are given statistically. These include the weather, the fertility
of the crops, and other factors of the sort. In addition there are certain
factors which we can control. For example, the amount of seed grain to be
planted, the rate of interest on agricultural loans, etc. The problem of
planning is to optimize, or in other words minimize, some quantity depending on the controllable and the uncontrollable statistical factors in such a
way that this minimization is maintained on the average. Such a problem is
of a statistical character, and therefore of an informational character.
From the economic point of view, this is what we may call the
statistically stable planning problem. If we have any planning situation
which is meant to continue, it must of necessity lead to a statistically stable
planning situation. On the other hand, for a planning situation to be statistically stable, it is not necessarily true that we can arrive at it from existing
286
19
287
and his direct style all help.5 Wiener's brief articulation on this subject
ignores details and points to a more lasting upshot of the bureaucratic
process. He invoked the "Circumlocation Office" from the novel Little
Dorrit by Charles Dickens. (This Office gives the run-around to Daniel
Doyce, the innovative and honest craftsman who wants to get a patent
[or his invention.) The reference to Little Dorrit occurs in the course of
Wiener's discussion o[ patent law [50j, pp. 113-115], but it serves to
show up one of the more lasting drawbacks of the bureaucratic process,
viz. its stifling of initiative. This strong anti-homeostatic factor is ingrained in existing systems of public administration.
The mechanical engineer prescribes definite forms of lubrication
for his machinery to compensate for friction, and the communication
engineer incorporates elaborate filters in his amplifiers to dampen noise.
The public manager, however, leaves equally significant dissipating
factors in the economic and political process unattended. Everyone
knows that the execution of a major economic or political undertaking
can be distorted by bureaucratic dissipation. Concerning such dissipation is a world literature ranging from the Hindu book of animal fables,
the Hitopedesa (500 B. C.), and the treatise, the Arthasastra (300 B. C.),
to the present day humorous writings of Parkinson. Yet as of now no
social scientist has tried to represent this persistent and significant factor
in his equations and flow diagrams. The measurement of bureaucratic
dissipation, and the design of filters to eradicate it, is an outstanding
problem in modern political economy that continues to get pushed
under the carpet.
While the games mentioned in the previous paragraphs are
genuine, there exists a tendency to see game-playing where there is none.
For instance, in his 1976 book, D. Bell writes:
Life in pre-industrial societies ... is primarily a ~ame against nature.
Industrial societies ... playa game againstjilbricated nature . ...
A post-industrial socicty ... is a game between persons. {B4, pp. 147, 148:
(author's italics)6
Very important events in the life of early man, such as the discoveries
offire, the wheel and writing, were not games at all but results of inquiry
5
His occasional humorous use of symbolism, for instance, IJ + J 5 -> IJ5' for
the bureaucratic-chemical reaction in which the elements, incompetence I,
and jealousy J, when mixed in the right proportions, yield the compound
injelitance, IJ" {PI, p. I02}, adds to the cybernetician's delight.
Bell's classification, viz. pre-industrial, industrial and post-industriaL is less
intrinsic than Wiener's classification: zeroth, first, second, ... industrial
revolutions, based on whether the tools of production exemplify the amplification of force. power, intelligence, ... , cf. 19A.
288
19
devoid of stratagem, cf. 18G. The stochastic nature of the cosmos does
not turn natural processes into games. The uncertainties of weather, for
instance, do not make the struggle offarming a game. Nor is the process
of manufacturing a car a game against anything. The only players
against whom man can play games are the animals whom he domesticates or kills or pushes around, and other men with whom he has conflicting relations. (Friendly games such as bridge are another matter.)
The latter conflicts, social, economic or political, were as much a part of
the life of early man as they are of modern life.
F Communal information
The existence of Social Science is based on the ability to treat a social group
as an organization and not as an agglomeration. Communication is the
center that makes organizations. Communication alone enables a group to
think together, and to act together.
What is true for the unity of a group of people, is equally true for
the individual integrity of each person. The various elements which make up
each personality are in continual communication with each other and affect
each other through control mechanisms which themselves have the nature of
communication. N. Wiener, cf. {D2, p. 77}
Of all these homeostatic factors in society, the control of the means of
communication is the most effective and the most important. [61c, p. 160]
Wiener was speaking of societies, too large for direct contact between
individuals, for which the press (books and newspapers), the radio, the
television, the theater, the schools and the church are essential for
communication. The last is what holds a community together, its cohesiveness being dependent on the amount of easily retrievable communal
information. Information is communal, if it so affects the interrelationship between two or more members, that this minority behavioral
change is noticed by the rest of the community, and spreads to the
majority. See [61c, pp. 157, 158).
Wiener pointed out that information is not a commodity, i.e. an
entity easily passable from hand to hand without loss of value, and that
this value is not an extensive (i.e. finitely additive) quantity such as the
volume or the weight [50j, p. 116]. In a closed system information
decreases spontaneously (Entropy Law). It is not storable without
"overwhelming degeneration of its value" and is useful only when it can
freely circulate, and assist the growth of knowledge [50j, p. 120).
It is dangerous, Wiener claimed, to leave the apportioning of the
channels of communication to the anti-homeostatic mechanism of the
market. For then, the ownership of a channel will fall in the hands of
those who operate it for its secondary benefits, viz. monetary gain, and
289
not for its primary benefit: the furtherance of socially conducive communal information. "The man who pays the piper, calls the tune." The
channels will be cluttered by the repetition of half-true, sex-symbolic
blurbs; in short, with misinformation. Worse still, as we saw in 19C,
the channel-owners will have a stake in the maintenance of mass gullibility and in the continuation of an indifferent educational set-up that
caters to it, cf. [61c, pp. 161, 162]. The result will be not informative
communication, but noise and misinformation; not homeostasis, but
conflict.
On the other hand, the considerably greater homeostasis of the
small communities such as those in New England, where Wiener grew
up, and which he came to love, cf. [53h, Ch. VII], stems from the
long-time sharing of uncontaminated communal information, and the
uniform levels of intelligence, behavior and altruism that this engenders.
In such a community:
The average man is quite reasonably intelligent concerning subjects which
come to his direct attention and quite reasonably altruistic in matters of
public benefit or private suffering which are brought before his own eyes. In
a small country community which has been running long enough to have
developed somewhat uniform levels of intelligence and behavior, there is a
very respectable standard of care for the unfortunate, of administration of
roads and other public facilities, of tolerance for those who have offended
once or twice against society ... in such a community, it does not do for a
man to have the habit of overreaching his neighbors. There are ways of
making him feel the weight of public opinion. [61c, p. 160]
In these communities the situation is not all favorable for the "merchants
The complexity in the social fields, to which Wiener refers, comes from
the strong coupling therein between the observer and the observed, and
from the proximity of the periods of their natural rhythms.
The methods of precise science have worked only in fields
characterized by a very loose coupling of observer and observed, i.e. in
which observation does not significantly disturb the observed object.
Our astronomical observations have no influence on astronomical
290
19
291
Here Wiener felt that the appropriate economic analysis had to embody
the fractal concept of Mandelbrot. Dissipative factors also pose a problem in the social fields as we saw in 19E.
Such considerations, rooted in the concepts of complementarity,
time scale discrepancy and the persistency of dissipation ( 19E), led
Wiener to reject systematic social engineering as a possible solution to
our social ills. Such engineering demands more precision and a looser
coupling between the social engineer and his object than is possible in
the social fields. Recall that Wiener had learned the Complementarity
Principle the hard way, while grappling with the difficulties of
generalized harmonic analysis and quantum mechanics in the late 1920s,
and that his awareness of time scales and chaotic time series, acquired
from his studies in the early 1920s, underwent severe testing in his war
work (Ch. 14). Wiener's thoughts on sociology thus bore the impress of
ideas he had acquired in the course of his deepest and hardest mathematical research.
292
19
For Wiener, like most scientists, believed in the unity of theory and
practice.
While Wiener was not acquainted with Marx's theories of surplus value and its expropriation, he was aware of the exploitation and
alienation of labor. { 19A, C} He knew that technological advances
such as the introduction of gas and electric lighting had led to the
imposition of night work in stuffy factories, and often worsened the
worker's life. To the question asked at the turn of the century by the
French Marxist, Karl's son-in-law, Paul Lafargue:
When Science subdued the forces of nature to the service of man, ought she
not to have given leisure to the workers that they might develop themselves
physically and intellectually; ought she not to have changed the "vale of
tears" into a dwelling place of peaee and joy? I ask you, has not Science failed
in her mission of emancipation? {Ll, pp. 25, 26)
293
not forget what he had learned from Perrin that "the logic of the
mathematicians has kept them nearer to reality than the practical representations of the physicists".
Furthermore, Wiener could not accept the dialectic materialist
position according to which the "paramount question of the whole of
philosophy" is: "Which is primary, Spirit or Nature?" (Engels {E7,
p. 21 and that "the answers which philosophers gave to the question
split them into two great camps" {E7, p. 2l}. To Wiener and Ashby, this
question, far from being paramount, was largely irrelevant. The line
between spirit and nature is fugitive as the great Marxist, J. B. S. Haldane understood, cf. 10D. The ontological differentiation, idealism
and materialism, is inconsequential as Wiener noted:
n,
In short,
dialectic materialism
dialectic spiritualism.
The first requirement for a sound philosophy is that it shake off the
shackles of ontology, and seek a Leibnizian universalism.
As to the actual implementation of dialectical materialism in the
Soviet Union, Wiener could hardly accept as enlightened a philosophical position that condemned the Principia Mathematica till the 1960s,
settled for a pseudo-scientific hodgepodge in genetics in the 1950s, and
called cybernetics "reactionary" until 1960-a position, moreover, that
tcnded to confuse the wisdom of the myth with superstition, and had
difficulty imbibing peaks of human endeavour such as the epics of
Homer, the Divine Comedy of Dante, the Paradise Lost of Milton,
Michelangelo's non-secular frescoes and Bach's St. Matthew Passion.
Wiener was not gullible enough to accept all that Marxism had come to
mean, and was skeptical of those who did so uncritically, and this
included most communists. Today, of course, many communists acknowledge the distortions their movement underwent under the fire of
experience. The solution, however, lies not in a glasnost (openness)
applied to message and noise alike, but in a tchestnost (honesty) that
recognizes the efficacious aspects of Pythagorean-Platonism.
Professor Heims's description of Wiener's attitude on the cold
war (involving some words of Struik {SlO, p. 37}) as being one of "a
plague on both your houses", cf. {H6, p. 311} is grossly inadequate.
Wiener understood the anti-homeostasis latent in capitalism ( 19C).
But he could hardly hail a political economy that trailed in advanced
technology and automatization, suffered food shortages, and had to
keep importing food and advanced technology from capitalist markets.
294
19
Nor could Wiener swallow the claim that the cold war represented a
class struggle between oppressors and oppressed, when what was hampering the struggle of the oppressed was dissension among communists
and chronic bureaucratism. Thus he had little to choose between the two
"isms". However, Wiener had great respect for the humane aspects of
Russian Socialism-such as free housing for the poor-as he had for the
freedom and democracy he loved in American towns. He was a humanist. If he ever declared blight on the two houses, it could only have been
on the bureaucratic diseases that have corrupted their high ideals.
Wiener had an abiding love for American democracy and was
ardently patriotic. He tells us that from the start "he was repelled by the
totalitarianism of the communists", but recognized that "the appeal of
communism" felt by his younger colleagues "was an appeal to their
humanitarian instincts" [56g, pp.220-22I]. With the humane side of
Marxism he had no quarrel. His objections to other aspects of Marxism
came not from adherence to a fixed liberal democratic position, but from
plain clearheadedness and objectivity, and from an honest disbelief in
groups that are afraid to trim the beards of their favorite prophets.
295
20
296
20
ber 1963, and the thank-you letter he received from the DeputyCommander, Major-General T. R. Stoughton, was profuse:
Your lecture was exceptionally well received and there have been many
favorable comments regarding it. The students were highly impressed with
your enthusiasm. the vitality and comprehensiveness of your presentation
and your candid responses to their questions.
We are most grateful that you found time in your busy schedule to
address the College, and we consider ourselves honored to have had the
benefit of your wide experience. [MC, 330]
The wrong impression that Wiener was hostile towards the U.S. military
after the war came from the attention attracted by his exaggerated
verbal reactions, and by his two immature articulations [47b, 48d], and
too little attention to his more mature judgments. The reactions
sometimes took the form of his turning down requests for de-classified
reports of his defense work, on the presumption that they would be used
to make new weapons of destruction. Such was his response, for instance, to requests from Dr. G.E. Forsyth (of December 3, 1946 [MC,
72] and Professor Oskar Morgenstern (of January 3, 1947 [MC, 75]). In
his replies Wiener pointed out that the reports could be procured
directly from Washington.
The reply to Forsyth had more, however, and its content was
reprinted in the Atlantic Monthly and the Bulletin of Atomic Scientists,
cf. [47b]. In it Wiener spoke of "the tragic insolence of the military
mind" and wrote:
I do not expect to publish any future work of mine, which may do damage
in the hands of irresponsible militarists. [47b]
These are among the worst of Wiener's articulations. We are left in the
dark as to who these "military minds" and "militarists" are. Are they
primarily the civilians who decide defense policy at the highest levels, or
are they senior officers in uniform who advise on the implementation of
policy?
The story of the atomic bomb shows the difficulty. Let us
begin at the beginning. What initiated the Manhattan Project was a
letter from Einstein and Szilard to President Roosevelt, and the recommendation to drop the atomic bomb without warning came from a
committee (appointed in spring 1945 by H.L. Stimson, the U.S. Secretary of War) that was headed by Wiener's distinguished colleague and
friend Dr. Vannevar Bush. It was accepted by President Turman, whose
military career was confined to World War I, and who can hardly be
described as a "military mind". This decision had been opposed by his
chief of staff, the Fleet Admiral W.D. Leahy, a "military mind", but
supported by General George C. Marshall, another "military mind".
Before the first Soviet nuclear explosion in 1949, when the option of
297
preventive nuclear strikes on the Soviet Union was in the air, among the
supporters of such strikes were scientists of the caliber of J. von Neumann (cf. Blair {BI2}), and Wiener's teacher, Bertrand Russell {R8,
vol. III, pp. 7, 8}, a World War I pacifist and, all told, one of the humane
philosophers of this century. Perhaps "the tragic insolence of some
civilian minds", or "the impudence of wealth and arrogance of power",
would fit the facts a little better. Wiener's article has several such
ambiguities. Some of these positions he reiterated in another short, but
noisy article [48d]. (These articles as well as an exposure of their misleading nature appear in the Call. Works, IV.)
In the early 1950s, however, Wiener the philosopher took over.
He sensed the futility of the position he had taken:
I tried to see where my duties led me, and ifby any chance I ought to exercise
a right of personal secrecy parallel to the right of governmental secrecy
assumed in high quarters, suppressing my ideas and the work I had done.
After toying with the notion for some time, I came to the conclusion
that this was impossible, for the ideas which I possessed belonged to the times
rather than to myself If r had been able to suppress every word of what I had
done, they were bound to reappear in the work of other people, very possibly
in a form in which the philosophic significance and thc social dangers would
be stressed less_ r could not get off the back of this bronco, so there was
nothing for me to do but to ride it. [56g, p. 308] (emphasis added)
298
20
Here Wiener asserts that (1) the football game concept of war is silly;
(2) such a concept is popular in our army and navy; (3) of the two effects
of dropping the atomic bomb-reduced American casualties, and the
commencement of a future conflict-the latter is just as important.
The assertion (1) is supported by the doctrine in General Karl
von Clausewitz's 1832 classic on war {CIO}i, and today it should be
obvious to a competent global strategist that modern wars, unlike
football games, form one type of weave that intermittently appears
within the growing fabric of world history. The assertion (2) was true
when Wiener wrote, and still is, it would seem. As for (3), the fact that
the United States chose to annihilate a couple of cities to save its troops
from combat triggered off a set of military responses from other powers.
The costliness of future struggles that might ensue from these responses
should have been carefully weighed in making the decision to exercise
the atomic option. Indeed, a responsible body of opinion (including
Admiral Leahy, Chief of Staff to Presidents Roosevelt and Truman) has
argued that a frontal attack on the mainland could have been averted,
American casualties significantly reduced, and the civilian population
spared, by a sustained naval blockade and by strategic air-strikes solely
against industrial targets. Captain Sir Basil Liddell Hart has pointed
out that by July 1945, the United States was made aware of Japan's
desire to end the war, by the appointment of a new government headed
by Prince Konoye {L9, pp. 694-695}. The atomic bomb was dropped on
August 6.
In short, if we ignore the tempests of the late 1940s, Wiener's
judgments reveal a good grasp of military realities.
Another of Wiener's serious concerns was the destructive effect
of war on scientific morale:
Before the war, and particularly during the depression, positions in science
were not easy to get. The requirements for these positions had become
exceedingly high. During the war, this situation had changed in two respects.
First, there were not enough men to carry out all the scientific projects which
the war involved. Secondly, in order to carry out these projects at all, it
became necessary to organize the work so as to use those with a minimum
amount of training, ability, and devotion.
The result was that young men who should have been thinking of
preparing themselves in a long-time way for their careers lived in a lightSee Note on pp. 316--317.
299
hearted way from hand to mouth, confident that the existing boom in
scientists had come to stay. Such men were in no state to accept the discipline
or hard work, and they evaluated whatever intellectual promise they might
have as if it had been already realized in performance. With the older men
crying out for assistance and manpower, these boys would shop around for
those masters who would demand least and grant them the most in indulgence and flattery.
This was a part of a general breakdown of the decencies in science
which continues to the present day. In most previous times, the personnel of
science had been seeded by the austerity of the work and the scantiness of the
pickings. There is a passage in Tennyson's "Northern Farmer: New Style"
which says "Doant thou marry for munny, but goa wheer munny is!" [56g,
p.27l]
Wiener also believed that more harm than good came from insistence on
undue secrecy in defense work, since it tended to lower quality by
inhibiting the flow of ideas as well as the flow of criticism. His belief was
backed by his experiences in his own defense work in World War II, cf.
14G.
There is nothing in all this that is "anti-militaristic" or to which
a good battle commander could object. No commander worth his salt
would settle for troops with a "minimum amount of training, ability and
devotion", who "shop" for easy and flattering masters, or who would let
his channels of communication be clogged by bureaucratic interference.
As for Wiener's non-association with and public criticism of the
leadership in Washington, this reflected a disapproval of policy, not of
the constitutional form of the American government nor of the Department of Defense per se. One can be pro-defense without being proPentagon, witness Admiral Hyman Rickover. Wiener had an abiding
love for American democracy-especially as it affected the small town,
cf. Ch. 3, and was ever ready to rally to its defense, as we know from his
conduct during the two wars.
Wiener's views on atomic bomb diplomacy and his iconoclastic
views on the American economy ( 19C) were not in tune with those of
the Truman and Eisenhower regimes. Finding official channels uncongenial, Wiener decided to devote himself to alerting the public to the
socio-economic problems that the impending age of automatization
would pose, and this led to his contacts with Walter Reuther, as we saw
( 19A). With the advent of the Korean War in 1950, however, Reuther's
attention shifted to geopolitical issues, and Wiener's followed suit.
B Wiener's letter to Walter Reuther on geopolitics, and von
Clausewitz's principles on war
300
20
301
American "knowhow", it is our own special way of bearing our breasts and scaring
our enemies, like the French "panache" and the British "doggedness". The quality is
there, but it is not an exclusive American possession.
6. As to the atomic bomb, it is in the long run our greatest liability. Asia
associates our use of it in Japan with a point of view in which yellow women and
children are worth less than white ones, and such soldier terms as "gook" do nothing
to dissipate this impression. To use the bomb again will put the seal of permanency
on our growing reputation for stupid brutality. It makes no difference that the
Russians are delighted for us to have acquired this reputation. This is all the greater
reason for not playing into their hands. I say that the use of the atomic bomb in the
last war was the work of a fool, and that its repeated use can only be the work of an
enemy of the United States.
7. Apart from this, from the purcly technical point of view, we have overreached ourselves in trusting to the bomb as a weapon. As industrial countries go, we
are old, and have established an economic system in which a vital part of our factory
potential is placed in our great cities: in exactly those places where the first killing
effect of an atomic bomb would be supplemented by the inevitable secondary disorganization and panic which it will produce. Our very military authorities, in expressing their confidence in our defenses, explicitly do not undertake to give us the
assurance that no Russian atomic bomb will land on its target.
8. Unlike our own, the Russian industrial potential has come late to the
scene, and has been developed in an era in which mass bombing, if not atomic
bombing, has been continually contemplated. An appreciable part of it has been
developed under the potential threat of the atomic bomb. In a country of vast
distances, it is unthinkable that this new order has identified the centers of industry
with the great cities to the extent to which we have done it. Thus in comparing the
Russian atomic strength with ours, it is a vast underestimation of the Russian
situation to count bomb against bomb.
9. Even now, a full-scale third world war will be for us a fight against odds.
The first conclusion is that if it comes, we should do nothing to hasten it, and that our
civilian defenses need as much bolstering as our military defense. Strictly speaking, as
long as we have not decentralized our cities, we are not on a war footing, no matter
how strong our armed forces arc. Nevertheless, this decentralization is so long and
expensive a job that it is not practical politics in less than a decade or so. The
conclusion is obvious: if we havc to fight at thc periphcry of our strcngth, let us do
so, let us gather strength, and lct us abide the issue. However, let us do nothing to
hurry on Armageddon, hoping, as is very probably true, that the Russians do not yet
carc to involve themsclves in the protracted and mutually destructive conflict that this
would be.
10. The preparation for war and its expenditures mean at least a slowing up
of social progress, and perhaps a reversal in its tide. We must not forget that there are
elements in this country which regard this slowing up, and this reversal, with sardonic
glee. It is the chance of a certain type of business man, and of a certain type of military
man, to get rid once for all of the labor unions, of all forms of socialization, and of
all restrictions to individual profiteering from below. It is a trend which may easily be
turncd into fascism. This we do at our own peril, for it is only thc personal and moral
advantage of the American way oflife over what the Soviets can offer us, which makes
it worth while for the average man, the body of the country, to undergo the hardships
and dangers of years of prospective conflict, rather than to surrender at once.
II. Even though we can look to the democratic countries only for a limited
immediate military help, it is far from a matter of indifference to cultivate their good
will, as well as the ultimate good will of countries which are at present Soviet satellites,
302
20
by offering them substantive financial and organizational help, as well as the even
more important intangible help which consists in the support of their developed and
inchoate liberal institutions. Certainly there is nothing to be said for the sabotage
which we have applied to all countries which like England, have accepted any form
of socialism. We must also avoid the backing of discredited and _2 regimes, such as
that of Chiang Kai Shek in China. We must show enough cooperative interest in the
problems of economically backward and primitive countries, such as India or China,
to make these countries feel that we can give them a more promising and more secure
future than can Russia. We must be alert to all plans in Europe to lay aside old
hatreds, and to unite.
12. If we do these things, and do not merely attempt to "sell" America to
Europe by a propaganda as shallow, as stupid, and as lying as any cigarette advertising campaign, we shall find a strong underground to work for us, not only in the
countries at present neutral, but even in some of the Russian satellites. We shall
eventually find divisions of soldiers rallying to our aid, and we shall have a good
fighting chance to survive in a world fit for us to live in-for this is the maximum of
what winning can mean with modern weapons. If however we fail to realize that we
can win the world only by accepting as ours the interests of the world, moral as well
as material, then we shall perish, as we shall deserve to perish.
13. This is my comment. Use it as you will, provided that if you publish it,
you keep it essentially together, and that you consult me.
Sincerely yours,
Norbert Wiener
One of the important principles in General Karl von Clausewitz's classic on warfare {Cl O}, asserts that the "most decisive act of the
statesman or general is to understand the kind of war in which he is
engaging, and not to take it for something else"3. In paragraphs 2, 3, 4
of his letter to Reuther, Wiener (who probably never read Clausewitz)
devotes himself to an exposition of "the kind of war" that the United
States would be facing in the 1950s. Wiener's letter is Clausewitzian in
other respects. Absent in toto is the episodic interpretation of war as a
"game", which starts when a gun is fired and ends when a flag marked
"victory" or "defeat" is waved. War is seen in both its military and
political phases as an ongoing contest. Wiener's perspective is global,
and does not (as in bad military planning) leave out or misread salient
factors such as the needs, attitudes and aspirations of the populaces and
the state of their morale. This Clausewitzian tone of the letter is not
surprising, for some of Clausewitz's ideas, such as war being a trinity,
one term of which, to wit its chaotic aspect, "makes it a free activity of
the soul" come rather close to Wiener's ideas, such as the dependence of
freedom on contingency. As the letter shows, Wiener instinctively veered
2 Illegible word.
3 Cf. Note on pp. 316-317, and the principle No.9. To appreciate what
follows, the reader should glance through the list of principles.
303
304
20
Chess playing automaton, built by Tores y Quevedo in 1890. Dr. Quevedo's son
and Norbert Wiener at the Cybernetical Conference in Paris, 1951.
any of the other forms of competitive activity in which we are really interested. [60d, pp. 1355, 1356] (emphasis added)
(3)
305
the adversary's theoretical shortcomings: his ignorance or prejudice or misunderstanding of history, and consequent inability to
foresee impending developments.
Hence, as Wiener put it:
... in determining policy ... there are several different levels of consideration
which correspond in a certain way to the different logical types of Bertrand
Russell. There is the level of tactics, the level of strategy, the level of the
general considerations which should have been weighed in determining this
strategy, the level in which the length of the relevant past-the past within
which these considerations may be valid-is taken into account, and so on.
Each new level demands a study of a much larger past than the previous one.
[60d, p. 1356]
306
20
For instance, during a battle within a global war, the soldiers and their
immediate officers make their move according to the von Neumann-Morgenstern theory. The generals behind the scene do not worry if the "board" keeps
getting worse and worse in a certain sector during a certain period, if this is
in accord with the overall war plan.
307
308
20
catenated transducers operating at different speeds. "Operating on different time-scales" was Wiener's expression. He believed, as we saw in
16F, that many an active transducer (e.g. the firefly) has its own
internal rhythm, and that this is controlled by its own internal "clock".
The rhythm of this clock determines its "time-scale". Wiener was wont
to analyze in these terms many a conflict, which most of us would
consider in ad hoc ways. It is worth noting that he addressed in this way
the very question of science and the promotion of human welfare that
engaged Professor Heims {H6}. Wiener's thoughts on this question are
articulated in the last section ("time-scales") of his address [60d] to the
AAAS. He poses the issue as follows:
What are the moral problems when man as an individual operates in connection with the controlled process of a much slower time scale, such as a
portion of political history or-our main subject of inquiry--the development of science? [60d, p. 1356)
In practical terms this would mean devoting a fraction of his time and
energy to the kind of societal thinking that the members of the Federation of Atomic Scientists engaged in.
What would Wiener's suggestion amount to if in the last paragraph we replace "the scientific enterprise" by "Clausewitzian war" in
the sense of an ongoing contest, and replace "scientist's activity" by
"execution of defense"? It would mean a smaller allocation of defense
departmental funds to weapons development and production, and a
larger allocation for the cybernetical analysis of weaponry and global
strategy in broad and far-sighted terms. This substitution is of course
309
2.
3.
4.
5.
6.
7.
8.
War is not a "glorified football game" which starts with the firing
of a gun and ends with the waving of a flag, but one phase of an
ongoing geopolitical process.
The wisdom of a military act depends on its immediate effects as
well as on its long-range ramifications on future contests.
Moves, tactics, strategies, superstrategies ... form a hierarchy of
types 0, 1,2,3, ... (a tactic is a sequence of moves, a strategy a
sequence of tactics, and so on). As in set-theory, there will be
trouble if type lines are uncritically transgressed.
The von Neumann-Morgcnstern game theory applies only in
military operations of low type n.
Great generals win by going up the type-ladder. Their moves are
based not on an evaluation of the board (as in checkers), but on
strategic evaluations that involve not just the enemy's limitations
in manpower and material, but also in his experience, military
know-how and political understanding.
Each new level demands a study of a much larger past than the
previous one.
Game theory is useful for low type military operations. At the
higher levels, a time-series analysis of the enemy's experience is
necessary.
For a concatenation of transducers to function smoothly, there
must be a free flow and feedback of information between them
310
9.
20
10.
11.
311
differences to bear on his study of how the unlimited growth ofscience and
technology might jeopardize the continuation of a moral social order. He
writes:
The traditional ethics of the scientific community, dictated by the god of
scientific and technological progress, needs modification so as to allow for
the reemergence of suppressed human and cultural needs. {H6, p. l59}
312
6.
7.
8.
9.
10.
20
313
government would furnish free insurance to American corporations up to half a billion dollars per nuclear accident.
These ten pieces of evidence are indicative of a clear bellicosity
in von Neumann's geopolitical position vis-a-vis Wiener's. Until those,
who have had the good fortune of being the direct beneficiaries of von
Neumann's wisdom and of knowing his true thought, can provide
contravening evidence, we have to proceed on the basis of the factuality
of this position. This writer requested Professor Abraham Taub, the
editor of von Neumann's Collected Works {V6} and author of his
biography {Tl}, to shed further light on the ten pieces of evidence. But
all that Professor Taub could suggest, by way of a possible refutation,
was von Neumann's testimony at the Oppenheimer security hearings. It
is von Neumann's geopolitical evaluations, however, that concern us,
not his stance on Dr. Oppenheimer's patriotism.
Unfortunately, we have very little from von Neumann's own pen
to tell us of the strategic thinking that underlay his belligerent stance. So
once again we are left to seek guidance from the little that is available.
A passage from a letter von Neumann wrote to Lewis Strauss in 1951
sheds some light on his thinking:
The preliminaries of war are to some extent a mutually self-excitory process,
where the actions of either side stimulate the actions of the other side. These
then react back on the first side and cause him to go further than he did "one
round earlier", etc .... Each one must systematically interpret the other's
reactions to his aggression, and this, after several rounds of amplification,
finally leads to "total" conflict. .. I think, in particular, that the USA-USSR
conflict will probably lead to an armed "total" collision, and that a maximum
rate of armament is therefore imperative. {H6, p. 287}
314
20
Adding that Engels's writing in the military field (over 2000 closely
printed pages) outnumbered those in any other field, he continues: "I
find much more to praise than to condemn in the specifically military
thinking of the great Marxists" {Gl, p. 67}
But von Neumann did not appreciate the enemy's strong points.
In his testimony before Congress in January 1955 (during confirmation
hearings pending his appointment as Commissioner of Atomic Energy),
he described himself as being "violently opposed to Marxism ever since
I can remember". There is no evidence to indicate that his opposition to
Marxism rested on a study rather than on hearsay or emotion. Von
Neumann did not know, for instance, that the ideas in his very important 1945 paper on general economic equilibrium {V5} had been anticipated by Karl Marx in the 1860s. This has been pointed out by the
eminent econometrist, Professor M. Morishima {M 17}.
These limitations of von Neumann's appreciation of the USAUSSR conflict in disregard of the dichotomous nature of war, cast doubt
on his maturity as a global strategist. Characterizations of him as an
"exclusively rational" strategist (Ulam {UI, p. 6}), or "hard boiled"
strategist (Blair {BI2}), are thus inaccurate. In truth, his idealization of
the cold war was scientifically inadequate, in stark contrast to the
profundity of his earlier idealizations of quantum phenomena and of
ratiocinating automata, and his astute though tentative idealizations of
non-linear phenomena. The bellicosity in his stance thus stemmed from
an insufficiently scientific appreciation of war, and not, as Professor
Heims {H6} suggests, from an exclusively scientific one.
The truth is that the excessive inhumaneness in our technological
society stems not from too much of a scientific attitude, but from too
315
little of it. The continuation of a moral social order, far from being in
conflict with the growth of science, as Heims's words (p. 311) suggest,
demands an evermore vigorous and widespread application of the scientific methodology. This was Wiener's position.
As to von Neumann's strategic stance, let us remember that his
official responsibilities were not in global strategy but in weapons development. He wrote no papers on global stratagem that we know of,
and his thoughts on the question do not affect his greatness as a scientist.
The fact that he was not an "exclusively rational" strategist is not
relevant.
A reader may nevertheless ask why von Neumann, a great
mathematician and scientist, instead of generalizing the theory of games
into a comprehensive theory of war stratagem, adopted a less than
scientific position on the matter. Why did Galileo, the father of modern
mechanics, slip on the question as to why water rises only up to 34 feet
in a pump? Why did H. Poincare, universal mathematical genius,
pathetically classify Cantor's Mengenlehre as a disease? Why did
P.E.A. Lenard, Nobel Prize-winning physicist, vent hocus-pocus about
the superiority of "Aryan science" over "Jewish science"? "The most
complicated object under the sun" was von Neumann's apt description
of the human nervous system (p. 243). This "complicated object" has
turned man into Homo sapiens, but only to a degree; it has also allowed
him to turn into something else. (See 2IB.)
G The Black Mass, twentieth century
Wiener's great analogical mind saw a parallel between the contemporary
concern over the misuse of technology and the medieval concern over
the misuse of the Mass-the so-called Black Mass. In this Mass the
orthodox dogma that "the priest performs a real miracle and that the
Element of the Host becomes the very Blood and Body of Christ"
remains the principle, but because of its ulterior motivation, the ceremony is severely condemned by the Church. To quote Wiener:
The orthodox Christian and the sorcerer agree that after the miracle
of the consecration of the Host is performed, the Divine Elements are
capable of performing further miracles. They agree moreover that the miracle of transubstantiation can be performed only by a duly ordained priest.
Furthermore, they agree that such a priest can never lose the power to
perform the miracle, though ifhe is unfrocked he performs it at the sure peril
of damnation.
Undcr these postulates, what is more natural than that some soul,
damned but ingenious, should have hit upon the idea of laying his hold on
thc magic Host and asking its powers for his personal advantage. It is here,
and not in any ungodly orgies, that the central sin of the Black Mass consists.
316
20
The magic of the Host is intrinsically good: its perversion to other ends than
the Greater Glory of God is a deadly sin. [64e, pp. 50-51] (emphasis added)
Wiener was aware of mankind's perennial apprehension of inherent evil in "the use of great power for base ends". It is the moral of
the story, "Fisherman and the Jinni" in the Arabian Nights, and of
several myths and legends of antiquity. The theme appears in Goethe's
The Sorcerer's Apprentice, and in other fiction, among which Wiener's
favorite was The Monkey's Paw, by W.W. Jacobs, cf. [64e, pp. 55-59].
In this way, Wiener was able to mobilize the "accumulated common
sense of humanity" to warn against the light-hearted use of the great
powers that modern science and technology have placed at our disposal
for base ends, i.e. for catering to the new sins and idolatries of the
modern secular age: materialism, gadget-worship and the transference
of the power of being beyond good and evil, which in truth is ideal and
belongs to God alone, to our all too human rulers.
Note: Von Clausewitz's principles of war
317
2.
3.
4.
5.
6.
7.
8.
9.
10.
318
21
Wiener's Excursion
into the Religious Domain
319
320
21
321
Prior to the Fall man had the potency to sin (i.e. to commit evil)
as well as the potency not to; he possessed the freedom to choose;
The Fall came from Adam's sin: "by desisting what was better
he committed an evil act" {A7, Ch. 34};
After the Fall, man lost his (natural) potency not to sin. 1
322
21
323
could easily intrude into the settlement of disputes within the tribe itself.
In this way the range of murderousness and deception would keep
widening, with increasing sorrow, despair and anxiety following in the
wake.
What has emerged from this evolution, as we know, is the most
self-centered and cruel of the mammalian species, a race made up of
self-alienated individuals. This retrogression is what we shall call the
Fall, for it is a scientific accounting of the kind or moral devolution of
man that the sages and saints wanted to portray. Briefly, the Fall is the
transition that took the hominid from the animal-hunt to the man-hunt,
and that brought in self-alienation.
This Fall differs from the process of phylogenetic corruption of
reproducing automata only in that the acquired characteristics are
transmitted by social heredity and not by Wiener's Lamarckian incubator ( 18F). The offspring acquire their latent skills in the violent
arts by early schooling from their parents. This Fall differs from Augustine's only in its substitution of evolutionary anthropogenesis in place of
monogenesis. This train of thought, resulting from Wiener's observation
of the analogy between entropy and evil, and his learning theory has two
positive aspects:
324
(1)
(2)
21
Wiener was aware that in man's struggle for emancipation from sin, the
faculties of imagination and idealization are crucial.
The importance of imagination for man's liberation from ignorance, avarice and injustice is obvious. Any movement away from the
status quo requires some conception of what "ought to be", as P. Tillich
{T4} has emphasized, and this last, since it does not exist, can only be
imagined. The origins of the faculty of imagination can be traced back
to animal ritual. This, as we now know, not only serves the emotional
needs of the species, but also joins with the external environment in
orienting its evolutionary trend.
Ritual becomes an even more vital formative agent for man. As
A. N. Whitehead has remarked: "Mankind became artists in ritual"
{W9, p. 21}. In the human ritual speech plays an important part. A rite,
i.e. a ceremonial ritual with incantation, invariably accompanies the
initiation of important activity such as the hunt. Myths, involving gods
and deities, are invented to lend credence to the rite, and so justify the
efficacy of the activity. These myth-makers were our first scientists, who
tried to explain why things work. In these myths each situation is perceived as a blurred organic whole, in much the way a child or a poet
perceives, incognizant of the distinctions and polarities occurring in a
325
Third, Wiener was aware of the challenging role of mythical and paradoxical concepts in science such as Maxwell's demon, cf. [50j,
pp. 28-30]. (The exorcising of this demon helped not only in the understanding of the modern statistical theory of information, but of the
lovely "demonic" role of chlorophyll particles, 12G.)
A typically Wienerian insight on the issue of science and myth
pertains to the fact that in mythology the name of the object is deemed
to have the potency of the object itself. (Examples: prayer, with expressions such as "we ask in the name of"; slogans such as "In the name of
Stalin, attack"Y Science has increasingly militated against such identification of symbol and essence, and indeed in areas such as metamathematics a sharp distinction between symbol and designatum, between
mention and use, is crucial. But in Wiener's theory of regeneration of
machines, the symbolic description of an object is made to serve as its
operative image, somewhat as in mythology. For instance, a good
diagram of an electric circuit can do the job of the circuit itself, for as
Wiener tells us:
... an electric circuit may fulfill a relatively complicated function, and its
image, as reproduced by a printing press using metallic inks, may itself
function as the circuit it represents. These printed circuits have obtained a
considerable vogue in the techniques of modern electrical engineering. [64e,
p. 31] (emphasis added)
Thus, the allied notions in mythology of (i) the name being a true image
of the object, and (ii) the name serving as the object (of the "word"
becoming "flesh" as it were), can be put to use in the designing of
efficient and inexpensive engineering hardware.
2 A slogan used by the Red Army during battles in World War II.
326
21
With the growth of the crafts, and the accompanying development of prose and logical discourse, imagination found an intellectual
arena. We have the idealizations of science, i.e. imaginative concepts of
ideal, non-existent entities suggested by entities that exist. A typical
example is Euclid's concept of a point, a being that has location but no
volume. Such idealizations are the starting points of scientific theorizing,
a fact amply illustrated in our previous chapters. Idealization, was, of
course, Wiener's forte.
Besides the mythic mode, a religious idea that Wiener found to
be of profound importance in science is faith:
I have said that science is impossible without faith. What I say about the need
for faith in science is equally true for a purely causative world and for one
in which probability rules. No amount of purely objective and disconnected
observation can show that probability is a valid notion. To put the same
statement in other language, the laws of induction in logic cannot be established inductively. Inductive logic, the logic of Bacon, is rather something on
which we can act than something which we can prove, and to act on it is a
supreme assertion of faith. It is in this connection that I must say that
Einstein's dictum concerning the directness of God) is itself a statement of
faith. Science is a way of life which can only flourish when men are free to have
faith. [50j, p. 193] (emphasis added)
327
Wiener spoke of the scholar as "the custodian of the intellectual development of this society, of the understanding of truths already known,
and of the development of new truths and concepts". He spoke of the
teacher as "the guardian of the task of passing down these new truths
and discoveries already known, to a new generation" [60e, p. 27]. Convinced of the great need of the scholar, the scientist, and the teacher for
phylogenetic survival, Wiener wrote:
... the tampering with the truth on the part of the man normally devoted to
it is a dereliction of duty quite comparable with the dereliction of an officer
who runs away from his own soldiers in the face of the enemy ... [60e, p. 27]
328
21
329
Perhaps the best thing that has been said about the continuity of the State
was formulated two centuries before Christ by the Chinese sage Menscius.
Menscius said, in effect, that the rule of the emperor is from heaven, but that
when a country has come through a long period of misrule and misfortune,
it is a sign that the emperor and even the dynasty has lost the mandate of
heaven, and that the country must seek elsewhere for its rulers. This view
represents an interesting attempt to combine a certain permanency in the
essentials of government with the transience of its details and in the selection
of those on whom the task of government lies. [62c, p. 36]4
Thus, when the ruling class has lost "the mandate of heaven", i.e. when
it holds on to an economic order that constricts the development of the
productive forces, the resulting social revolution destroys the ruling
class, but not the institution of the State.
Wiener brought his experiences in the prediction of time-series to
point to the great differences in the objectives, mode of inquiry and type
of prediction involved in long-time and short-time considerations. Even
in air warfare, a policy that has been viable over a six-month period
could prove disastrous if prolonged say for six years [62c, p. 32]. A
short-time type 1 policy involves type 1 feedback: the deviation of shell
from bomber (type 0 entities). For a long-time policy to succeed, it must
involve type n feedback: the deviation of the (n-l) type policy from its
goal.
In general, short-time planning may be based on reasonably
accurate forecasts, deterministic or stochastic. When it is the latter, only
the linear extrapolation of a short segment of a fairly stationary time
series is usually involved. Such prediction would be futile for long-term
purposes, say planning for a city like Athens with its 2,500-year history.
The extrapolation has now to take into account a much larger pastsegment of the time-series. The latter is non-stationary, and the prediction, to be useful, must be non-linear.
The use of this long-time information is so different from the one of shorttime information that it is not economical to trust them to the same instruments and the same computation. [62c, p. 32]
For instance, to finance long-time city planning by private investmentfor-profit would be about as absurd as extending a short-sprint technique to run a marathon race (cf. 19C). Briefly, "The support of the
long-time needs of the human race cannot be left exclusively on the basis
of the returns they make on the welfare of short-time institutions", [62c,
p. 36] (emphasis added).
4 It should be noted that a dual theory of kingship, similar to the Chinese,
prevailed in India at about the same time, 300 B.C. {CI2}, and that the
corresponding conception emerged in Europe, under a neo-Platonic auspices, in the two sword governmental ideas of St. Augustine {A8}.
330
21
331
332
21
reflection based on the destruction of Rome by Alaric. Naturally, Wiener could never share the Marxian belief in the actual removability of
alienation, i.e. in the actual transformation of civil society into communism, "an organization in which the free development of each is the
condition for the free development of all" {M8}. For him, communism,
like the Sermon on the Mount, offers an ideal to mankind, forever
unattainable but one nevertheless towards which it must forever strive.
Guided by the great tragedy writers of ancient Greece, Wiener
sought dignity in this tragic situation. The Second Law of Thermodynamics tells us that all life must come to an end.
Yet we may succeed in framing our values so that this temporary accident
of living existence, and this much more temporary accident of human existence may be taken as all-important positive values, notwithstanding their
fugitive character. [50j, p. 40]
333
Thus, in the light of Wiener's own analogy, man's quest for a personal
God appears as an attempt to adapt the Pythagorean-Platonic aspect of
7 Father Faramelli footnotes, however, that he got this information from
Wiener's family, friends and colleagues and not from Wiener's writings {Fl,
p. 328}. We would refer the interested reader to Father Faramelli's study {FI}
for other aspects of the religious implications of Wiener's thought, carried
out from a standpoint that is more ontological than ours. For more on the
methodological standpoint here adopted, see {MIl}.
8 Einstein, quoted in The New York Times, 25 April, 1929, p. 60, column 4.
334
21
Note. The landscape on page 323 does little justice to the calamity
it is meant to depict. Graphic pictures of the holocaust are available. See F. Greene, Let There Be a World, Fulton Publishing Co.,
Palo Alto, California 1963, or Hibakusha, translated by G. Sekimori, Kosei Publishing Co., Tokyo, 1984.
9 Cf. A.N. Whitehead: "Mathematics and the Good", {WID, pp. 666-681}.
335
22
336
22
337
338
22
It was not quite clear that this letter implied your request for
permission to translate my book into Armenian, but Professor Parseghian
visited me this morning and told me that you were still waiting for permission. You and I know that with the absence of a copyright on books
published outside the Soviet Union, such a permission is not necessary for
you to go ahead. On the other hand, I appreciate greatly your asking for such
a permission. I am thoroughly sympathetic with your project and should be
highly complimented if you should proceed with your plans for such a
translation.
In loyalty to my publishers over here, I must register a formal
protest concerning the absence of copyright for foreign books. This protest,
however, is purely formal as I know that nothing can be done about it.
Therefore, privately and without committing my publishers in any way, I say
"go ahead" with my best wishes. And do accept my apologies for not
answering sooner.
Sincerely yours,
Norbert Wiener
Wiener was not merely a reader of mystery novels and science fiction; he
also enjoyed watching them played out in the theater and in motion
picture. He especially liked plots that centered on paradoxes. l Not
surprisingly, Wiener wanted to see some of his own literary ideas put on
the stage or in motion picture. In such situations he sometimes let his
wishes be known to producers of plays and movies. The following letter
to Alfred Hitchcock is a case in point. After two unduly apologetic
paragraphs Wiener gets down to business [MC, 146]:
He once mentioned in this regard the movie The Bridge On the River Kwai.
In this, a British officer's actions, conforming to the best moral code of the
British army, and the actions of a British commando unit, just as dutiful,
combine in in a way that inexorably leads to catastrophe.
339
340
22
Wiener goes on to describe his typical day at MIT; his courses, his
contacts with engineers, young and old. He then comments on his
immediate circle of colleagues and the little use they have for "organized
research". After a philosophical discussion of how good research, both
pure and applied, is done, he charges de Kruifwith "misplaced emphasis"
in his portrayal of C. F. Kettering, and points to Oliver Heaviside as the
real discoverer of the distortionless telephone line. The letter ends with
the words:
Heaviside, by the way, should be rich literary material for you.
In this connection, I should like to see a series of biographies of
engineers from your hand. Other names could be Nicola Tesla and B.A. Behrend. I should be glad to render you any possible help in the matter.
341
I trust that you will not take my remarks amiss. If they interest you in any
way, I should be delighted to hear from you.
342
22
Our thesis is not that the arts are an expression of mathematics through the
senses, but that mathematics itself is in the strictest sense of the word, a fine
art. In this the author finds himself in complete accord with the views
expressed by Havelock Ellis in "The Dance of Life". [29h, p. 129]
343
me display it to the world." In any case it is not a matter of the artist laboring
to express the emotion to the public but rather of the emotion endeavoring to
express the artist.
The relation between the mathematician, his public, and his emotion
is quite the same. You may rationalize his thoughts into some such statement
as "this is an interesting idea. Let me show my colleagues-my public-where
it leads". This is quite similar to the rationalized statement of the artist's impulse, and like it, is quite false. The mathematician does research because the
research demands to be done. [29h, pp. 130, 131] (emphasis added)
This attitude towards the aesthetic impulse comes close to that of the
great scholastic thinkers, e.g. Meister Eckhart, when he wrote: "What I
say, springs up in me, then I pause in the idea, and thirdly I speak it out"
or Dante, when he said: "Who paints a figure, ifhe cannot be it, cannot
draw it", cf. A.K. Coomaraswamy {CII, pp. 7, 175-17S}. In his masterful treatment, Coomaraswamy has summed up the scholastic view of the
aesthetic intuition in the words:
Whatever object may be the artist's chosen or appointed theme becomes for
the time-being the single object of his attention and devotion; and only when
the theme has thus become for him an immediate experience can it be stated
authoritatively from knowledge. {ell, p. 7}.
(2) In his article on Aesthetics [ISb] in the Encyclopedia Americana, written in 1917, Wiener deals at length with its history, starting with
344
22
Plato's The Idea of the Good. But he dismisses the period from Plotinus
to the Renaissance in the single sentence:
... although we do find treatments of aesthetic problems from the time of
St. Augustine to that of St. Thomas, these have been singularly arid and
without fruit in modern aesthetic theory. [18b, p. 199]
Thus the mathematics, which is a fine art, includes not only the work of
G. H. Hardy but also that of Einstein. Whether a mathematical inquiry
is initiated to settle a point of physics (in the wide sense) or ofmathematics is of little consequence. What is of considerable consequence is
whether the abstraction resulting from the inquiry is aesthetically appealing and rich enough to sustain a deductive system of sufficiently
wide scope.
3 Unfortunately Wiener did not revise his article on Aesthetics, and so the 1960
edition of the Encyclopedia carries the same ignorant remark on the scholastic period as the 1917 edition,
345
[29h] with one written forty years later with a title which Wiener could
have used for his own, viz. Mathematics as a Creative Art, by P. R.
Halmos {H3}. In this the conclusion reached in the last paragraph is
flatly denied. Halmos writes:
As I sec it the main difference between mathophysics and mathology is the
purpose of the intellectual curiosity that motivated the work. .. {H3, p. 384}
So far so good-all would agree that Hardy and Einstein had different
purposes. But then Halmos adds:
The mathophysicist wants to know the facts, and he has, sometimes at any
rate, no patience for the hair-splitting pedantry of the mathologist's rigor
(which he derides as rigor mortis). The mathologist wants to understand the
ideas, and hc placcs grcat value on the aesthetic aspects of the understanding
and the way that understanding is arrived at; he uses words such as "elegant"
to describe a proof. In motivation, in purpose, frequently in method, and
almost always, in tastc, the mathophysicist and the mathologist differ. {H3,
p. 385}
Here we run into difficulty. In the first place, "rigor" becomes "rigor
mortis" not just for our Diracs and Eddingtons, but also for our Eulers,
Ramanujans, Galois and other such mathologists. Second, and more
important, the habit of placing "great value on the aesthetic aspects of
the understanding and the way that understanding is arrived at" extends
well beyond the circle of mathologists, and certainly includes the great
mathophysicists. For in their mathophysical quests, Pythagoras, Archimedes, Kepler, Newton, Lagrange, Hamilton, Clerk Maxwell, Einstein,
de Broglie, Schrodinger, Weyl, Wiener and von Neumann, among others, were very much alive to the aesthetic origins and aesthetic import of
their ideas, some seeing in them reflections of "celestial harmony", "the
order of Nature" and "Spinoza's God". And much of the most sublime
mathologistic architecture is founded on conceptions unearthed by these
aesthetically conscious mathophysicists.
We have therefore to reaffirm Wiener's conclusion in [29h] that
from the standpoint of aesthetics the division into mathology and
mathophysics is of no importance; good mathophysics is as much a
creative art as good mathology. And mediocre mathology can be as
unartistic as mediocre mathophysics.
But a moment's reflection suggests that these dicta must apply to
the entire spectrum of human labor. For if there is an aesthetic aspect
to the activity of a good mathophysicist, how can this not be the case for
the activities of a good experimental physicist or a good geologist or a
346
22
347
Only a contingent and noise infested cosmos has room for art
and creativity. In a strictly deterministic world these concepts disappear.
Creativity is criticality, i. e. the ability to tell message from noise, and
create message without producing more than minimal noise. Wiener, of
course, knew that an automaton could beat the human in an intellectually difficult game such as chess, and saw in the computer a prosthesis
for the human brain, cf. 20D, 16E. Nevertheless, he regarded criticality, the ability to spot message and desist from waste, to be a very
important constituent of intelligence, and he felt that automata were
4 Cf. Halmos {H3, p. 382}: "I don't think a team of little Gausses could have
obtained the theorem about regular polygons under the leadership of a rear
admiral any more than a team of little Shakespeares could have written
Hamlet under such conditions."
348
22
deficient in this regard. By virtue of their low impedance and high speed,
electronic automata avoid a lot of waste of energy and time, and so
possess an elemental criticality. With their present design, however, their
activity forces them to wade through heaps of balderdash; they thus fall
short of Wiener's standard. Wiener felt that intelligence can be arranged
in a hierarchy of types,s and that the type of the most intelligent of
present day automata is much lower than that of the human brain. He
was confident, however, that increasing knowledge of cerebration will
bring under construction intelligent automata of higher and higher
types. From the methodological standpoint, he maintained a scientific
posture and did not slip into the quagmire of "anti-reductionism".
5 The universality of the Turing machine does not affect this issue, Wiener
would have contended, since its universality does not enhance its criticality.
349
23
350
23
was a poor listener. His self-praise was playful, convincing, and never
offensive. He spoke many languages but was not easy to understand in any
of them. {FS, p. 344}2
Levinson has described his experiences as a student in Wiener's postgraduate course ("really a seminar course") in 1933:
He would actually carryon his research at the blackboard. As soon as I
displayed a slight comprehension of what he was doing, he handed me the
manuscript of Paley-Wiener for revision. I found a gap in a proof and proved
a lemma to set it right. Wiener thereupon sat down at his typewriter, typed
my lemma, affixed my name and sent it off to a journal. A prominent
professor does not often act as secretary for a young student. He convinced
me to change my course from electrical engineering to mathematics. He then
went to visit my parents, unschooled immigrant working people living in a
run-down ghetto community, to assure them about my future in mathematics. He came to see them a number of times during the next five years to
reassure them until he finally found a permanent position for me. (In those
depression years positions were very scarce). {L8, pp. 24-2S}
This little story is more telling of Wiener, the man, than the earlier
impression. But Levinson hastens to add:
If this picture of extreme kindness and generosity seems at odds with Wiener's behavior on other occasions, it is because Wiener was capable of
childlike egocentric immaturity on the one hand and extreme idealism and
generosity on the other. Similarly his mood could shift quickly from a state
of euphoria to the depths of dark despair. {L8, p. 2S}
351
When Wiener's name began to appear in newspapers, he started receiving letters from strangers asking all sorts of questions. A human being
reveals himself to an extent in the way he handles this situation. Mahatma Gandhi, for instance, made it a point to respond in some way to
every letter he received. A good portion of the 70 volumes of his
Collected Works {G2} comprise correspondence. One such letter, from
an Attica Prison inmate under life sentence for murder, was the following [MC, 49]:
P.O. Box 149
Attica, N.Y.
September 19, 1938
Dear Doctor:
In the September 14th issue of the New York Times I read of your
discovery of a new Calculus.
For the past seven years-during which time I have been in prison
-I have devoted all of my time to the study of calculus, including Finite
a six-week period in Bombay in 1953-1954. This led to collaboration at the
Indian Statistical Institute in Calcutta in 1955-1956 and at MIT in 19571958, and to intermittent exchange of ideas thereafter at MIT and at his
summer home in South Tamworth, New Hampshire.
352
23
There followed two more letters. The first from the prisoner, dated
September 25, 1938, begins with the words [MC, 49]:
Your letter dated September 23 was most inspiring. It is difficult to explain
in writing the grand reaction I experienced when your letter arrived explain-
353
ing your interest. After seven years imprisonment it is good to know that
someone takes an interest in my educational development.
The second letter from Wiener informed Scimone that Levinson and he
were sending him Hardy's Pure Mathematics, Goursat's Cours d'analyse and Osgood's Lehrbuch der Funktionentheorie, [MC, 49, September 30].
From the other end of the spectrum of strangers, Wiener received the following short letter [MC, 125]:
My dear Dr. Wiener,
I am a Victorian. I should like to know what kind of world my
grandson is going to live in. Will you kindly give me the names of some of
your books on the new industrial revolution-referred to in the enclosed
clipping. Some nice easy ones-pleasant to read.
Sincerely yours,
Ella J. Meyer
(Mrs. J. Franklin Meyer)
3727 Jocelyn St.
Washington, D.C.
September 21, 1950
We could not find out what reply, if any, Wiener made. 4 It is fair to
surmise that in this area too, Wiener's conduct could range from extreme concern to helpless indifference.
To turn to Wiener's relationships within the academic community, his most noticeable shortcoming was as a teacher-"a famously
bad lecturer" is Freudenthal's apt description {F5, p. 344}. Struik remarks how Wiener could "occasionally lull an audience to sleep", but
how he was also able to hold groups of colleagues and executives "at
breathless attention while he manipulated his ideas and flares of vision"
{SlO, p. 35}. Levinson spoke of him as "a most stimulating lecturer", but
at the seminar level {L8, p. 24}. Only very rarely, however, did Wiener's
lecturing attain these peaks of exultation or depression. For the most
part it hovered not too far from the minimum: it was chaotic, but
amusingly and not irrevocably so. A typical, off-par Wiener lecture may
be described in the words used by Freudenthal in a somewhat different
context:
After proving at length a fact that would be too easy if set as an exercise for
an intelligent sophomore, he would assume without proof a profound theorem that was seemingly unrelated to the preceding text, then continue with
a proof containing puzzling but irrelevant terms, next interrupt it with a
totally unrelated historical exposition, meanwhile quote something ... and
so on. {F5, p. 344}
4
354
23
355
356
23
chess by following a minimax technique. The idea was quite sound, but
Wiener never followed up with a paper on the designing or programming of such a machine. This was done by C.E. Shannon in 1950 {S8}.
Such writing, useful though it may be, is not well received in scientific
circles, especially when the author makes it a habit.
C Resignation from the National Academy of Sciences; muddled ideas
on science academies
357
and that the Academy is maintained by dues. However, the fact remains that the
Academy is accustomed to regard itself as a government agency, to ask diplomatic
privileges for its official representatives in their travels, and in other ways to speak as
the scientific mouthpiece of the United States of America. As such, we have a
government agency which bestows titles of honor; which is based on the principle of
superordination and subordination, not of functions in an organization, but of
personalities; and which in many other ways is in glaring contradiction with the
declared principles of the United States of America as well as the actual practice in
which these principles are embodied in the government at large.
With the second function of thc Acadcmy, that of the custodian of certain
journals and funds for research, I have no quarrel, providing that the Academy
accepts a position simply on the same level as that of other agencies with a like
custodianship. I have no sympathy whatever with the Science Fund idea, which seems
to me an excellent means to discourage independent gifts to science, and to stifle all
work not pleasing to whatever group is at the moment running the scientific politics
of the country. I say this with full respect to the personnel now in charge of the fund.
When they go, the overcentralization of scientific funds will remain.
As to the third purpose of the society-the conveying of honors-I have no
sympathy at all. I have always regarded exclusiveness as an attribute chiefly of use in
selling unwanted junk to parvenus. I do not wish to belong to any scientific organization which has more than one grade of membership, nor to one in which that grade
of membership is not available to every person with a sincere interest in the field. We
all judge the ability of others, but I have no desire to see my unsolicited opinion of
a man published with official sanction to injure either him or his competitors, nor will
I accept such an unsolicited and officially published opinion of myself nor of anyone
else. This would apply to the best available opinion, from which, either because of
organized electioneering by influential colleges, government departments, and commercial laboratories, or because of the general fallibility of a group of persons none
too well-informed concerning the work of one another, I have found the official
judgment of the Academy to differ quite appreciably. As a young man, I have felt far
too much of the weight of the unsolicited disapproval or sanction of the elders of
science to wish to have any connection with a body of self-appointed judges. Every
time a new member is appointed, an unnecessary gift of prestige or position is made
to one man; and this gift comes from the one place from which it can come: from the
pockets and reputation of someone at a more remote institution or with less influential friends. I am afraid that I can not be reconciled to injustice even by becoming its
beneficiary.
As to medals, prizes, and the like, the less said of them the better. The
heartbreak to the unsuccessful competitors is only equalled by the injury which their
receipt can wreak on a weak or vain personality, or the irony of their reception by an
aging scholar long after all good which they can do is gone. I say, justly or unjustly
administered, they are an abomination, and should be abolished without exception.
So long as I am a member of the American Mathematical Society, I shall work against
the acceptance of a single penny or gift to be spent on medals or prizes, and for the
liquidation of those prize funds already established. I can not in honor continue in an
organization devoted in principle to their support.
I do not wish to speak in detail of the many faults I have had to find with
the Academy-of the bad catering, of the tedious and expensive dinners, of the
general atmosphere of select and costly pomposity which has hung over the meetings,
of the camp-followers of the press and the camera, of the excessive age of most of the
new members-first, because you have taken strong steps to improve these situations;
and secondly, because they do not touch the essence of my attitude, which is, that I
358
23
This letter, despite the generosity of its tone and its memorable
words, "I cannot be reconciled to injustice even by becoming its beneficiary" has to be grouped with Wiener's less mature writings. For apart
from the logically fuzzy treatment of the notions of "official", "quasiofficial", "responsible", etc., Wiener does not bring the full force of his
tremendous understanding to bear on the important issues he raises in
the letter, and does not propose amendments to the charter of the
Academy; instead he offers his resignation.
Even though Wiener's cybernetical ideas had not crystallized in
1941, he surely had an inkling (of what he was to say in 1961) that the
chief function of science is
that of subs erving a homeostasis in human life; that of maintaining a rapport
with the environment, which will enable us to face our environment and its
changes, as we may come to them. 5
One aspect of this is finding and making the best environment for man
as a physiological unit, a long-time problem in Wiener's terminology.
Another is grappling with emerging social problems such as overpopulation, economic dislocation and war. Wiener knew that to fulfill these
functions, there had to be semi-official bodies of scientists that could
advise governments on scientific matters. He knew of the origins of the
Accademia dei Lincei (1603), the Royal Society (1662) and the Academie Francaise (1666) and their singular service in the development of
science and in the formulation of governmental policy. Wiener criticizes
the system of appointment to the Academy by co-option by the academicians, and raises the question of accountability, but he offers no
alternative arrangement. In fact, the quest for alternatives runs us into
a quagmire, as is apparent from the following quotation of what
359
Briefly, we are confronted with the fact that when it comes to assessment
of merit and to voting, scientists are as fickle as the rest of mankind/
and that there is no obvious solution to the selection problem that is
clearly superior to co-option.
There is a bit of irony in Wiener's condemnation of the honorific
role of the Academy and its bestowal of prizes. For speaking of his
receipt of the Bocher prize of the American Mathematical Society in
1933, Wiener recounts that "it was a pleasant thing to be recognized ... "
[56g, p. 177]. And after his resignation Wiener did not decline honorary
degrees from Tufts College, 1946, and Grinnell College, 1957, nor the
Alverega Prize from the College of Physicians in Philadelphia, 1952, nor
the Medal of the Rudolf Virchow Medical Society, 1957, nor the
ASTME Research Medal, 1960, nor the National Medal of Science,
1964. We also know of the excellent research that was stimulated by the
prizes offered by the Academie Fran<;aise. The frenzy for virtuosity,
priority and recognition is a post-Renaissance aberration, encouraged
by secularism and by a social order which bungles on the hope that greed
can be made to stimulate useful activity. Wiener was by no means
6 Bernal does not say how the class of "qualified scientists" is to be defined.
7 See C.P. Snow's novel The Masters {S9} for this fickleness and the tragicomic consequences.
360
23
immune to the frenzy, as the Kellogg episode (Ch. 8) testifies. But until
a substantial change occurs in the socio-economic and moral climate,
science will have to cope with those drawn to its temple for the pleasures
of brain sport or for reasons of ambition, money-making, or vanity.
In his autobiography Wiener describes the National Academy of
Sciences and his connection with it in the following terms:
This is the organization which was entrusted during the Civil War with the
task of putting the services of the scientists of the United States at the
command of the American government. In the course of the years. its
governmental importance had gradually given way to the secondary function
of naming those American scientists who might be considered to have
arrived. There has always been a great deal of internal politics about science,
and this has been distasteful to me. The building of the National Academy
was for me a fit symbol of smug pretentiousness, of scholarship in shapely
frock coat and striped trousers. After a brief period, during which my
inquisitiveness concerning the nature of the high brass of science was amply
satisfied, I got out. [56g, p. 176]
From this one might surmise that Wiener's resignation from the Academy was caused by its internal politics rather than by dissatisfaction
with the more fundamental aspects of its organization that he voiced in
his resignation letter. When questioned about this, Wiener mentioned
that he became tired of the canvassing practiced by the "Harvard
group". He cited G.D. Birkhoff's pressure on him to back the election
to the Academy of the late Professor J.L. Walsh. Wiener would have
probably done this on his own, for he held Professor Walsh in high
regard, but he resented the pressure. This "pressure" could only have
come in 1936 when Wiener was in China. In his letter to Wiener of
January 8, 1936, [MC, 44], G. D. Birkhoff pointed out that Walsh had
received the highest number of votes cast, and added:
This is just enough to enahle us to present Walsh's name and I hope you will
join me in making a vigorous effort to see that he goes through. To this end,
would you write me a letter, which may be used as seems best, concerning
your estimate of Walsh's work? What you write will, of course, have much
influence.
Birkhoff's letter is not very different from the kind that a church trustee
might write to another who is away. And it is hard to see how any
organization can function without such preliminary agreements. But
Wiener disapproved of such requests, and evidently by 1941 had gotten
tired of them.
In this episode Wiener-noise seems to have beaten the Wienermessage. For apart from the ambiguities in the letter, it does not give the
full reasons for his dissatisfaction. Dr. Jewett's reply of September 24
(delayed by illness) ends both sensibly and sympathetically [MC, 60]:
361
While I still feel you are making a mistake and that you can render a better
service by staying inside the Academy and using your influence to make it
conform more nearly to what you think it should be, I realize that you alone
must judge your own desires.
I am sorry I have not been able to dig up a problem which would
show you the value I see in a body like the Academy even though it is not
all I myself should like to have it. However, one cannot always produce white
rabbits out of a hat on demand.
Whatever your final decision, believe me to be,
Sincerely your friend,
Frank B. Jewett
President
362
23
363
364
23
Norbert Wiener at the White House with President Lyndon B. Johnson and colleagues
(among them Drs. Weisner and Bush) when he received the National Medal of
Science in 1963.
365
Fortunately, Wiener did not let his prejudices interfere with his
scholarly pursuits. He imbibed freely what Harvard had to offer, and
made Birkhoff's theorem his own. Nor, barring one or two exceptions,
did he allow his subjective attitudes to affect his scientific criticism. His
prejudicial statements should therefore be overlooked as the mild aberrations of a great and fundamentally impartial and honest mind.
366
24
Epilogue
24 Epilogue
367
... our one hope probably lies in atomic energy ... A pound of matter has
11,300,000,000 kilowatt hours of energy in it, while a pound of gasoline as
used in an internal combustion engine has at most about one kilowatt hour.
The ratio is so enormous that if even a modest part ... of the energy in
matter can be made available for mechanical use, all other forms of storage
will be put far in the shade.
Mr. Robert 1. Van de Graaff, of the Institute's Department of Physics ... is
quite reasonably optimistic as to the ultimate possibility of realizing, mechanically at least, a large part of the energy which appears as a change of
mass of the nucleus when one element is synthesized into others, as in the
work of Cockcroft and Walton. [33f, pp. 59, 70]
He also of course understood thoroughly the practicality of good abstraction, as is clear from the applications he made of the Lebesgue
integral in Brownian motion theory and electrical engineering, and the
type-classification he brought into automata theory. Unlike his distinguished contemporary E.P. Wigner, who was puzzled by what he perceived as "the unreasonable effectiveness of mathematics in the natural
sciences", cf. {W13}, Wiener found the effectiveness of mathematics
perfectly understandable. The investigations of the remarkable medical
men close to him, W. Ross Ashby, A. Rosenblueth and W. McCulloch,
368
24
Epilogue
Wiener not only shared this belief, but, in Struik's words, he "lived the
unity of science" {SIO, p. 34}.
Wiener did not let the revolutions of thought that make and
partition the history of science obscure his vision of its fundamental
continuity. The age of science began when man first made an attempt to
comprehend the world around him, and this happened when his mode
of perception and language were still poetic. "Mouths of rivers", "veins
of minerals", such were his first scientific utterances. Wiener knew that
his own cybernetical ideas went back at least to Leibniz, and that
cybernetical devices were used from very early times. We now know that
Plato had barely missed the notion of feedback. The notion of fractal
too has roots extending to Aristotle, as Mandelbrot has indicated {M3,
p. 406}. Unlike some modern writers, Wiener did not forget that nonEuclidean geometry, Mengenlehre and fractals are inconceivable without the work of Euclid, and that there was nothing "anti-Euclidean"
about them. To him, as to J.E. Littlewood, the Greeks were "dons from
another college", who would have fully approved of such later activity.
In his view, the stark separation in the arts between the old and the new
is entirely absent in the sciences.
A similar perception of continuity marked Wiener's vision of
history as a whole. An admirer of both the 16th-century Renaissance
and the 18th-century French Enlightenment, he was blinded by neither.
The removal of medieval teleology from post-Renaissance science was
a boon, but Wiener contributed to its useful restoration in a modern
scientific setting. The same applies to his restoration of the long-time
State, the "Sacerdotium'. He was a revolutionary-traditionalist in the
best sense of the word.
Wiener was too great a thinker to accept uncritically all the
mores of post-Renaissance science. He denied Locke's classification of
"rational self-interest" as a virtue, and the belief in "unlimited progress"
24 Epilogue
369
370
24
Epilogue
371
1894
1895
1901
1903
1906
1909
1910
1911
1912
1913
1914
1915
1916
372
1917
1918
1919
1920
1924
1925
1926
1928
1929
1931-1932
1933
1934
1935
1936
1937
1938
1940
1935-1936.
Attended the International Congress of Mathematicians in Oslo, Norway, and lectured on Tauberian gap theorems.
Collaborated with Harry Ray Pitt at MIT during 1936-1937.
Delivered the Dohme lecture at Johns Hopkins University on
Tauberian theorems.
Lectured on analysis at the semicentennial of the AMS.
Appointed chief consultant in the field of mechanical and electrical aids
to computation for the National Defense Research Committee.
1941
1945
1946-1950
1946
1947
1948
1949
1950
1951
1952
1953
1953~1954
1955~1956
1956
1957
1959
1960
373
Consultant with the NDRC's Office of Scientific Research and Development, Statistical Research Group and Operational Research Laboratory at Columbia University.
Consultant to the War-Preparedness Committee of the American
Mathematical Society.
Joined a team at MIT under S. H. Caldwell to study the guidance and
control of antiaircraft fire.
Worked on the theory and design of fire control apparatus for antiaircraft guns with Julian Bigelow, under NDRC Project.
Resigned from the National Academy of Sciences.
Participated in a study group set up by John von Neumann, and
attended a meeting on communication theory in Princeton.
Collaborated with Arturo Rosenblueth at the Instituto National Cardiologia in Mexico, and attended the Mexican Mathematical Society's
Conference held in Guadalajara.
With Arturo Rosenblueth received a five-year Rockefeller Foundation
grant that allowed them to collaborate in Mexico and at MIT on
alternating years.
Received an honorary Sc.D. degree from Tufts College.
Attended the first three Josiah Macy, Jr. Foundation Conferences and
the Conference on Teleological Mechanisms sponsored by the New
York Academy of Sciences.
Lectured at the National University of Mexico.
Visited England and France, and gave lectures on harmonic analysis in
Nancy, France.
Spoke at the AMS's Second Symposium on Applied Mathematics.
Received the Lord & Taylor American Design Award.
Delivered the AMS's Josiah Willard Gibbs Lecture at the annual
meeting.
Attended the seventh Macy conference.
Lectured at the International Congress of Mathematicians at Harvard
University.
Lectured at the University of Paris, College de France, under a Fulbright Teaching Fellowship, and also lectured in Madrid.
Received an honorary Sc.D. degree from the University of Mexico.
Received the Alvarega Prize from the College of Physicians III
Philadelphia.
Delivered the Forbes-Hawks Lectures at the University of Miami.
Lectured on the theory of prediction at the University of California at
Los Angeles.
Taught a summer school course with Claude Shannon and Robert
Fano on the mathematical problems of communications theory.
Lectured at the Tata Institute of Fundamental Research, Bombay,
attended the All-India Science Congress, and visited research centers.
Visiting professor at the Indian Statistical Institute in Calcutta.
Lectured in Japan and gave a summer school course at UCLA.
Received an honorary Sc.D. degree from Grinnell College.
Awarded the Virchow Medal from the RudolfVirchow Medical Society.
Gave a summer school course at UCLA.
Appointed institute professor at MIT.
Lectured at the University of Naples in Italy, and visited the USSR.
374
1961
1962
1963
1964
1964
375
Shikao Ikehara
Sebastian Littauer
Dorothy W. Weeks
James G. Estes
Norman Levinson
Henry Malin
Bernard Friedman
Brockway McMillan
Abe M. Gelbart
Donald G. Brennan
Ph.D.
Sc.D.
Ph.D.
Ph.D.
Sc.D.
Ph.D.
Ph.D.
Ph.D.
Ph.D.
Ph.D.
1930
1930
1930
1933
1935
1935
1936
1939
1940
1959
1 Reprinted with the kind permission of Professor Irving Ezra Segal of MIT.
376
Mathematical Papers
A Mathematical philosophy and foundations
B Potential theory
C Brownian movement, Wiener integrals, ergodic and chaos theories, turbulence and statistical mechanics
D Generalized harmonic analysis and Tauberian theory
E Classical harmonic and complex analysis (orthogonal developments,
quasi-analyticity, gap theorems, and Fourier transforms in the complex
domain)
F Hopf-Wiener integral equations
G Prediction and filtering
H Relativity and quantum theories
I Miscellaneous mathematical papers
II Cybernetical and Philosophical Papers
A Philosophical papers
B Cybernetical papers
III Social, Ethical, Educational, and Literary papers
IV Book Reviews, Prefaces, and Obituaries
A Book reviews and prefaces
B Obituaries
V Abstracts
VI Books and Other Publications.
In the accompanying Bibliography, Wiener's publications have been classified into six
broad categories labeled I, ... , VI, and these larger categories have been divided into
subcategories labeled A, B, C, ...
The publications are indexed by the official year of their appearance. The
internal ordering of the publications appearing in a given year is not chronological but
according to the categories mentioned in the last paragraph: a run down the list a, b,
c, ... for a particular year entails a run down the list of categories lA, lB, ... , usually
with omissions and repetitions, as will be apparent from a glance at the right-hand
columns in the bibliography. For instance, [33d] means "the dth paper, according to
category, which appeared in a journal marked 1933".
References in the book indicated by numbers in square brackets preceded by
MC, e.g. [MC, 57], refer to the Manuscript Collection of Wiener in the MIT Archives.
Roman numerals in square brackets, e.g. [IV], refer to Defense Department Documents pertaining to Wiener, listed after the Bibliography. (Symbols in braces, e.g.
{K3}, which refer to other authors, are listed at the end of the book.)
377
[13a]
[l4a]
[14b]
[l4c]
[14d]
[15a]
[15b]
[16a]
[16b]
[16c]
[16d]
[l7a]
[17b]
[ISa]
[ISb]
[ISc]
[lSd]
[lSe]
I Sf]
IA
IA
IA
IIA
IIA
,IA
IA
I, I
IVA
IVA
lA
IVA
IVA
IIA
I, I
IIA
ITA
IIA
378
[18g]
[18h]
[18i]
[18j]
[19a]
[19b]
[19c]
[19d]
[1ge]
(19f]
(19g]
[19h]
[20a]
[20b]
[20c]
[20d]
[20e]
[20f]
[20g]
[20h]
[20i]
[20j]
[21a]
[21 b]
[21c]
[21d]
IIA
IIA
IIA
IIA
I, I
IIA
IIA
IIA
IIA
IIA
IIA
IIA
IA
IA
IA
IA
IA
IC
IVA
IIA
IIA
IIA
IA
IA
IC
IC
[22b]
[22c]
[22d]
[22e]
[23a]
[23b]
[23c]
[23d]
[23e]
[23f]
[23g]
[24a]
[24b]
[24c]
[24d]
[24e]
[24f]
[24g]
[24h]
[24i]
[25a]
[25b]
379
I, I
IA
IA
IA
IE
I, I
IA
IB
IB
IC
IC
ID
I, I
IB
IB
IB
IC
IC
ID
IVA
IVA
IVB
IB
ID
380
[25c]
[25d]
[25e]
[25f]
[26a]
[26b]
[26c]
[26d]
[26e]
[26f]
[27a]
[27b]
[27c]
[27d]
[27e]
[27f]
[27g]
[27h]
[27i]
[27j]
[27k]
[28a]
[28b]
[28c]
ID
ID
IE
I, I
ID
ID
ID
IH
IH
I, I
ID
ID
ID
ID
IE
IH
IH
IH
IH
I, T
I, T
ID
ID
IH
381
IH
ID
ID
ID
IE
IH
IH
IH
IIA
IIIE
IVA
ID
IVA
IF
I, I
IIIE
ID
ID
IIA
IIIE
IVA
V
IC
ID
ID
ID
382
[33e]
[33f]
[33g]
[33h]
[33i]
[34a]
[34b]
[34c]
[34d]
[34e]
[35a]
[35b]
[35c]
[35d]
[35e]
[35f]
[36a]
[36b]
[36c]
[36d]
[36e]
[36f]
[36g]
[37a]
[37b]
IE
IlIA
IVA
IVB
VI
IC
IE
IIA
VI
IIID
IE
IIIA
IIIC
IIIC
V
IIID
IE
IE
IE
IE
IE
I, I
IIA
ID
[3ge]
[39t]
[39g]
[39h]
[40a]
[40b]
[40c]
[40d]
[41a]
[41b]
[42a]
[43a]
383
IC
ID
ID
ID
ID
IE
IIIE
IVA
IC
IC
ID
IE
ID
IVA
IVA
V
IVA
IVA
IVA
V
IC
IC
IE
IC
384
[43b)
[44a)
[45a)
[45b)
[46a)
[46b)
[47a)
[47b)
[48a)
[48 b)
[48c)
[48d)
[48e)
[48f]
[49a)
[49b)
[49c)
[49d)
[4ge)
IIA
IVA
IG
IIA
IF
IIB
IE
IIIB
IIB
IIB
IIB
IIIB
IVA
VI
IG
IIB
IIB
IIB
IIB
[49t]
[49g]
[49h]
[50a]
[50b]
[50c]
[50d]
[50e]
[50t]
[50g]
[50h]
[50i]
[50j]
[50k]
[5Ia]
[51 b]
[5Ic]
[52a]
[52b]
[52c]
[53a]
385
IVB
VI
iVA
ID
TG
IIA
IIA
lIB
lIB
lIB
IlIA
lIlA
VI
VI
IIB
lIB
ITA
lIB
VI
lIB
IG
386
[53b]
[53c]
[53d]
[53e]
[53f]
[53g]
[53h]
[53i]
[53j]
[53k]
[54a]
[54b]
[54c]
[55a]
[55b]
[55c]
[55d]
[55e]
[56a]
[56b]
[56c]
IH
IH
lIB
IlB
lIB
lIB
VI
IlIA
IlIA
IlIA
lIB
I1IE
IlIA
IG
IG
IH
lIB
lIB
ID
IG
IH
[57b]
[57c]
[57d]
[57e]
[57f]
[57g]
[57h]
[58a]
[58b]
[5Sc]
[58d]
[SSe]
[58f]
[5Sg]
[5Sh]
387
IIB
IIB
IlIB
VI
IC
ID
ID
IG
IIB
IlIA
mc
IIB
IH
IG
IIB
IG
IG
IIB
JIB
IlIA
388
[58i]
[59a]
[59b]
[59c]
[59d]
[5ge]
[60a]
[60b]
[60c]
[60d]
[60e]
[60f]
[60g]
[61a]
[61b]
[61c]
[62a]
[62b]
VI
IG
IG
IG
IIIA
VI
IIB
IIB
lIB
IlIA
IIIB
IVA
IIB
JIB
JIB
VI
IIB
IIB
[62c]
[63a]
[63b]
[63c]
[64a]
[64b]
[64c]
[64d]
[64e]
[64t]
[65a]
[65b]
[65c]
[66a]
[66b]
[66c]
389
IlIA
IH
lIB
llID
lIB
lIB
lIlA
lIlA
VI
VI
IlB
IIB
IIB
IH
VI
VI
390
[75a]
[85a]
[85b]
[85e]
lIB
lIB
lIB
lIB
391
Ia. S.H. Caldwell, Proposal to Section D2, NDRC (3 p.), November 22, 1940.
lb. N. Wiener, Principles governing the construction of prediction and compensating
apparatus (8 p.) accompaniment to la, November 22, 1940.
II. K.T. Compton, Letter to Dr. Warren Weaver, NDRC, May 13, 1941.
III. J.H. Bigelow, Minutes of Conference held at Bell Laboratories on June 4, 1941.
IV. N. Wiener, Letter to Dr. Warren Weaver, NDRC, December I, 1941.
V. G. R. Stibitz, Note on prediction networks a la Wiener (14 p.), February 22, 1942.
VI. Warren Weaver, Letter to Dr. J.C. Boyce, MIT, March 24, 1942.
VII. N. Wiener, A.A. Directors, Summary Report of Demonstration (17 p.), June 10,
1942.
VIII. Demonstration by Wiener and Bigelow at MIT, July I, 1942, Diary of G.R.
Stibitz, Chairman, Division D2, July 23, 1942.
IX. N. Wiener: Final report on Section D2, Project No.6 (8 p.) submitted to Dr.
Warren Weaver, NDRC, December I, 1942.2
X. N. Wiener, Letter to Dr. Warren Weaver, NDRC, January 15, 1943.
XI. R. S. Phillips and P. R. Weiss, Theoretical calculation on best smoothing of position data for gunnery prediction, MIT Radiation Laboratory Report 532, February
16, 1944.
XII. N. Wiener, Automatic Control Techniques in Industry, Industrial College of the
Armed Forces, Washington, D.C., 1952-1953.
392
References
Al
References
393
B13
Bohm D., Causality and Chance in Modern Physics, Harper & Row, New
York,1957.
B14 Bohr, N., Atomic Physics and Human Knowledge, Wiley, New York, 1958.
B15 Bonaventura, St., De reductione artium ad theologiam. Translated by Sister
E. T. Healy, St. Bonaventure College, 1940.
B16 Born, M., Theory and Experiment in Physics, Cambridge University Press,
Cambridge, 1944.
B17 Born, M., The Natural Philosophy of Cause and Chance, Clarendon Press,
Oxford, 1949.
B18 Born, M. and Wolf, I., Principles of Optics, 5th ed., Pergamon Press, Oxford,
1975.
B19 Bre1ot, M., Norbert Wiener and potential theory, pp. 39-41, in Ref. A2.
B20 Bridgman, P.W., The Nature of Physical Theory, Princeton University Press,
Princeton, New Jersey, 1936.
B21 Buchdahl, H.A., The Concepts of Classical Thermodynamics, Cambridge
University Press, 1966.
B22 Bush, V., Operational Circuit Theory, Wiley, New York, 1929.
B23 Bush, V., The differential analyser-A new machine solving differential
equations, J. Franklin Institute, 212 (1931), 447-488.
B24 Bush, V., Gage, F.D. and Stewart, H.R., A continuous integraph, J. Franklin Institute, 203 (1927), 63-84.
C1
C2
C3
C4
C5
C6
C7
C8
C9
ClO
C11
C12
D1
D2
Cameron, R.H. and W. T. Martin, The orthogonal development of nonlinear functionals in series of Fourier-Hermite functionals, Ann. of Mathematics, 48 (1947), 385-392.
Cannon, J. W., Topological, Combinatorial and Geometric Fractals. Hedrick
Lectures of Math. Assoc. of Amer., 1982 (unpublished).
Carnap, R., The Logical Syntax of Language. Kegan Paul, Trench, Trubner,
London, 1937.
Carnap, R., Foundations of Logic and Mathematics, International Encyclopedia of Unified Science, Vol., I, No.3, University of Chicago Press,
Chicago, 1939.
Carnap, R., Introduction to Semantics, Harvard University Press, Cambridge, Massachusetts, 1942.
Carnap, R., Logical Foundations of Probability, University of Chicago Press,
Chicago, 1950.
Carnap, R., Philosophical Foundations of Physics, Basic Books Inc., New
York, 1966.
Cassirer, E., An Essay on Man, Yale University Press, New Haven, Connecticut, 1944.
Cassirer, E., Language and Myth, Dover Publications, New York, 1946.
Clausewitz, Karl von, On War, Translated by Colonel I. Graham, Routledge
& Kegan Paul, 1966.
Coomaraswamy, A.K., The Transformation of Nature in Art, Harvard University Press, Cambridge, Mass., 1935.
Coomaraswamy, A.K., Spiritual Authority and Temporal Power in the Indian
Theory of Government, American Oriental Society, New Haven, Connecticut, 1942.
Daniell, P.I., A general form of integral, Ann. of Math. (2) 19 (1917-1918),
279-304.
Deutsch, K., The Nerves of Government, Free Press, New York, 1966.
394
D3
D4
D5
D6
D7
EI
E2
E3
E4
E5
E6
E7
Fl
F2
F3
F4
F5
F6
GI
G2
G3
G4
G5
G6
G7
G8
References
Dewey, J., Logic, the Theory oj Inquiry, H. Holt & Co, New York, 1938.
Dewey, J., Education Today, Greenwood Press, New York, 1940.
Doob, J.L., Stochastic Processes, Wiley, New York; Chapman & Hall,
London, 1953.
Doob, J.L., Wiener's work in probability theory, pp. 69-72, in Ref. A2.
Draganescu, M., Odobleja Between Ampere and Wiener, Romanian Accad.
Sci., Bucharest, 1981.
Eddington, A. S., The Mathematical Theory oj Relativity, Cambridge University Press, Cambridge, 1937.
Einstein, A., Investigations on the Theory oJthe Brownian Movement, (1905),
Methuen, London, 1926.
Einstein, A., Methode pour la determination de valeurs statistiques d'observations concernant des grandeurs soumises a des fluctuations irregulieres.
Archive des Sciences Physiques et Naturelles, t. 37 (1914), p. 254-255.
Einstein, A., Sidelights on Relativity, Methuen, London, 1922.
Einstein, A., Out oj My Later Years, The Citadel Press, Secaucus, New
Jersey, 1950.
Eliot, T. S., Murder in the Cathedral, Faber & Faber, London, 1935.
Engels, F., Ludwig Feuerbach and the Outcome oj Classical German
Philosophy, (1888), International Publishers, New York, 1941.
Faramelli, N. Fr., Some Implications oJCyberneticsJor our Understanding oj
Man: An Appreciation oj the Work oj Dr. Norbert Wiener. Ph.D Thesis,
Temple University, (Department of Religion, 1968).
Fisher, R.A., Indeterminism and natural selection, Philosophy Sci. 1 (1934),
99-117.
Foures, Y. and I. E. Segal, Causality and analyticity, Trans. Amer. Math.
Soc. 98 (1955), 384-405.
Frank, P., Einstein, His Life and Times, Jonathan Cape, London, 1948.
Freudenthal, H., Norbert Wiener, pp. 344-347 in Dictionary oj Scientific
Biography, Vol. XIV, Chas. Scribner's Sons, New York, 1976.
Fuller, H.J., The Emperor's New Clothes: or Prius dementat, Scientific
MonthlY, LXXII (no. 1), January, 1951.
Gallie, W.B., Philosophers oj Peace and War, Cambridge University Press,
1978.
Gandhi, M.K., The Collected Works oj Mahatma Gandhi, Publications
Division, Ministry of Information and Broadcasting, Delhi, 1958.
Garcia Ramos, G., Libro homanaje: Arturo Rosenblueth, Instituto Politecnico Nacional, Mexico, 1971.
Giuculescu, A. The concepts of cybernetics. An historical outline,
pp. 139-204 in Ref. D7.
Gleason, A. M., Measures on the closed subspaces of a Hilbert space, J. Rat.
Mech. Analysis 6 (1957), 885-894.
Glushkov, V.M., Introduction to Cybernetics, Academic Press, New York
and London, 1966.
Glushkov, V.M., Contemporary cybernetics, pp. 47-70, in Survey oJCybernetics, Edited by J. Rose, Gordon & Breach, New York, 1969.
Godel, K., Uber formal unentscheidbare Siitze der Principia Mathematica
und verwandter Systeme 1, Monatsh. fiir Math. u. Phys. 38 (1931), 173-198.
References
395
G9
Kac, M., Wiener and integration in function space, pp. 52-68 in Ref. A2.
Kakutani, S., Two-dimensional Brownian motion and harmonic functions,
Proc. Imp. Acad. of Sciences, Tokyo 20 (1944), 706-714.
K3 Kakutani, S., Determination of the spectrum of the flow of a Brownian
motion, Proceedings, National Academy of Sciences, U.S.A. 36 (1950),
319-333.
K4 Keynes, J.M., Laisser-faire and Communism, New Republic, New York,
1926.
K5 Kolmogorov, A. N., Foundations of the Theory of Probability, (1933), Chelsea Publishing Company, New York, 1950.
K6 Kolmogorov, A. N., Sur l'interpolation et extrapolation de suites stationnaires, C.R. Acad. Sci. Paris, 208 (1939), 2043-2045.
K7 Kolmogorov, A.N., Wiener's spiral and certain other interesting curves in
Hilbert spaces, DAN SSSR, 26 (1940) 115-118. (In Russian.)
K8 Kolmogorov, A.N., Stationary sequences in Hilbert space (Russian). Bull.
Math. Univ., Moscou, 2, No.6 (1941), 40 pp. (English translation by Natasha
Artin).
396
K9
KI0
KII
Ll
L2
L3
L4
L5
L6
L7
L8
L9
LIO
Ml
References
Kolmogorov, A. N., Interpolation and extrapolation of stationary random
sequences, Izvestiya Akad. Nauk (Mathematical Series 5) (1941), 3-14.
Kosambi, D.D., Obituary: George David Birkhoff, Math. Stud. 12 (1945),
116-120.
Kratkii filosofskii slovar (Short philosophical dictionary), Edited by Mark
Moiseyevich Rosenthal, 4th ed., Leningrad, 1954, 236-237.
Lafargue, P., Socialism and the Intellectuals, New York Labor News, New
York 1967.
Lau, K. S., On the Banach space of functions with bounded upper means,
Pac. J. Math. 91 (1980), 153-172.
Lau, K. S., On some classes of Hardy spaces, J. Functional Analysis, (to
appear in 1989).
Lee Y. W., Synthesis of electric networks by means of Fourier transforms of
Laguerre's functions, J. Math. Phy 11 (1932),83-113.
Lee Y.W., Statistical Theory of Communication, Wiley, New York, 1960.
Lee, Y. W., Contributions of Norbert Wiener to linear theory and nonlinear
theory in engineering, pp. 17-34, in Selected Papers of Norbert Wiener, MIT
Press, Cambridge, Massachusetts, 1964.
Lenin, V.1., What Is To Be done? (1902), Edited by S.V. and Patricia Utechin,
Oxford University Press, Oxford, 1933.
Levinson, N., Wiener's Life, pp. 1-32 in Ref. A2.
Liddell Hart, B. H., History of the Second World War, Paragon Books, New
York, 1970.
Lynd, Albert, Quackery in the Public Schools, Grosset & Dunlap, New York,
1953.
References
M 14
MIS
M16
MI7
MI8
NI
N2
N3
N4
397
McCulloch, W., and Pitts, W., A logical calculus of the ideas immanent in
nervous activity, Bull. Math. Biophys. 5 (1943), 115-133.
Mehra, J. & Rechenberg, H., The Historical Development of Quantum
Theory, Vol. III, Springer-Verlag, New York, 1982.
Monod, J., Chance and Necessity: An Essay on the Nature Philosophy of
Modern Biology, Vintage Books, New York, 1972.
Morishima, M., Marx's Economics: A Dual Theory of Value and Growth,
Cambridge University Press, Cambridge, 1973.
Mumford, Lewis, Technics and Human Development, Harcourt, Brace
Jovanovich, New York, 1967.
Nagy, B. Sz., and C. Foias, Harmonic Ana(ysis of Operators on Hilbert Space,
North Holland, Amsterdam, 1970.
Nelson, E., Dynamical Theories of Brownian Motion, Princeton University
Press, Princeton, New Jersey, 1967.
Neurath, 0., Unified seiencc as encyclopedic integration, pp. 1-27, in International Encyclopedia of Unified Science, Vol. I, Chicago University Press,
Chicago, 1938.
Nordhoff, C., The Communistic Societies of the United States, (1875), Dover
Publications, New York, 1966.
01
PI
P2
P3
P4
P5
P6
QI
Q2
RI
R2
R3
R4
R5
R6
R7
398
References
R8
R9
SI
S2
S3
S4
S5
S6
S7
S8
S9
SIO
TI
T2
T3
T4
T5
T6
T7
T8
UI
U2
VI
V2
V3
V4
Russell, B., The Autobiography of Bertrand Russell, Vols. l, II, III, Little,
Brown, Boston, 1968.
Russell, B., Wisdom of the West, Crescent Books, New York, 1977.
Saeks, R., Causality in Hilbert space, SIAM Rev. 12, (1970), 357-383.
Samuel, A. L., Some studies in machine learning using the game of checkers,
I.B.M. Journal 3 (1959), 211-223.
Schetzen, M., Wiener and Volterra Series, Wiley-lnter-Science, New York,
1980.
Schrodinger, E., What is Life? Cambridge University Press, Cambridge,
1946.
Schuster, A. R., The periodogram of magnetic declination, Camb. Phil.
Trans. 18 (1900), pp. 107-135.
Shannon, C. E., A symbolic analysis of relay and switching circuits, Trans.
Am. Inst. Electr. Eng. 57 (1938), 713-723.
Shannon, C. E., The mathematical theory of communication, Bell Syst.
Techn. Journ. 27 (1948), 379-423; 623-656.
Shannon, C.E., Programming a computer for playing chess, Phil. Mag. 41
(1950), 246-275.
Snow, Sir C.P., The Masters, Macmillan, London, 1951.
Struik, D.J., Wiener: Colleague and Friend, American Dialogue, MarchApril, 1966.
Taub, A.H., John von Neumann, pp. 46-57, in Encyclopedia of Computer
Science and Technology, Edited by H. Belzer, A.G. Holzman and A. Kent,
Vol. 14, Marcel Dekker, New York, 1980.
Tawney, R. H., Religion and the Rise of Capitalism, (1926), Mentor Edition,
New York, 1954.
Taylor, G.!., Diffusion by continuous movements, Proc. Lond. Math. Soc.
20 (1920), 196-212.
Tillich, P., The Socialist Decision, Harper & Row, New York, 1977.
Toynbee, Arnold, Civilization on Trial, Oxford University Press, Oxford,
1948.
Tran, L. T., The Hausdorff dimension of the range of the N-parameter
Wiener process, Ann. Probability, 5, (1977), 235-242.
Tran, L. T., The Hausdorff a-dimensional measures of the level sets and the
graph of the N-parameter Wiener process, (1977), (unpublished).
Turing, A.M., On computable numbers, Proceedings, London Mathematical Society, Series 11, 42 (1937), 230-266.
Ulam, S., John von Neumann 1903-1957, Bull. Amer. Math. Soc., 64 (3),
Part II, May 1958, 1-49.
Ulam, S., Adventures of a Mathematician, Scribner, New York, 1976.
von Neumann, J., Eine Axiomatisierung der Mengenlehre, J. fur reine u.
angew. Math. 154 (1925), 219-240.
von Neumann, J., Mathematical Foundations of Quantum Mechanics (1930),
Princeton University Press, Princeton, N.J., 1955.
von Neumann, J., Physical applications of the ergodic hypothesis, Proc. Nat.
Acad. Sci., 18 (1932), 263-266.
von Neumann, J., Zur Operatorenmethode der klassischen Mechanik, Ann.
of Math. (2), 33 (1932), 587-642.
References
V5
V6
V7
WI
W2
W3
W4
W5
W6
W7
W8
W9
WIO
WI I
W12
WI3
WI4
Yl
399
Yaglom, A M., An Introduction to the Theory of Stationary Random Functions, Prentice-Hall, Englewood Cliffs, N.J., 1962.
Y2 Yaglom, A.M., Einstein's 1940 paper on the theory of irregularly fluctuating
sums of observations, Problemy Peredachi Informatsii (Problems ofInformation Transmission), 1985, Vol. 21, No.4, pp. 101-107. English translation:
IEEE ASSP Magazine, October 1987 (Vol. 4, no. 4).
400
Name Index
Name Index
Broad, C.D. (1877-1971): 72
Broglie de, L.V. (1892-1987): 26,119,
122, 123, 125, 129, 130,241,345
Brown, G.S. (b. 1907): 163,396
Brown, Robert (1773-1858): 81, 82
Browning, Robert (1812-1889): 30
Buchdahl, H.A. 157,393
Bush, Vannevar (1890--1974): 96, 101,
160--162,165,166,170-172,174,175,
179, 239, 241, 264, 295, 296, 366, 393
Butler, Nicholas Murray
(1862-1947): 61
Butler, Samuel (1835-1902): 35
Cairns, T. 205
Caldwell, Samuel H. (1904-1960): 181,
184,373,391
Cameron, R. H. (1908-1989): 151, 266,
271,393
Cannon, J.W. 85,393
Cannon, Walter B. (1871-1945): 34,168,
197, 198,203, 361
Cantor, Georg (1845-1918): 48,50, 59,
76,93
Caratheodory, C. (1873-1950): 94, 157
Carnap, R. (l87l-l970): 49,54,72,218,
223, 393, 394
Carnot, Sadi (1837-1894): 147, 148,254
Carroll, Lewis (1840-1921): 335
Cartwright, Mary (b. 1900): 136
Cassirer, E. (1874-1945): 25,326,394
Caton, R. 233
Chafetz, Morris E. 339
Chase, M. W. (b. 1905): 248
Chiang Kai-shek (1887-1975): 302
Christopherson, D.G. 180
Church, A. (b. 1903): 220
Clausewitz, Karl von (1780-1831): 298,
302-303,307,308,313,314,316,394
Cockcroft, Sir John D. (1897-1967): 367
Compton, Karl T. (1887-1945): 92, 192,
391
Conant, James Bryant (1893-1978): 361
Coomaraswamy, A.K. (1887-1947): 329,
343, 346, 394
Courant, Richard (1888-1972): 105
Cousins, Norman (b. 1912): 311
Cramer, Harald (1893-1985): 194
Croce, Benedetto (1866-1952): 136
Daniell, P.J. (1889-1946): 78,79,90,
394
Dante, Alighieri (1265-1321): 293,343
Darwin, Charles Robert
(1809-1882): 249
Davenport, H. (1907-1969): 47
Dean, Gordon 310
Dedekind, J.W.R. (1831-1916): 50
Deem, G. 151
De Kruif, Paul (1890--1971): 340,341
Delbruck, Max (1906-1981): 245,248
401
Democritus (b ab 460 B.c.): 143
Descartes, Rene (1596-1650): 277
Deutsch. Karl (b. 1912): 288,289,330,
336, 394
Dewey, John (1859-1952): 61,62,
277-279, 371, 394
Dickens, Charles (1812-1870): 287
Dirac, Paul A.M. (1902-1984): 119,260,
345, 368
Doob, J. L. (b. 1910): 83, 394
Draganescu, M. 253, 394
Duns Scotus, John (1265'1-1308): 62
Eckhart, Meister J. (l260?-1327): 343
Eddington, Sir Arthur (1882-1944): 26,
57, 152, 345, 394
Edwards, C. B. 71
Ehrenfest, P. (1880--1903): 147
Ehrenfest, T. 147
Einstein, Albert (1879-1955): 19,25,26,
47, 53, 55, 72, 79, 80-83, 85, 90, 102,
113, 115, 120, 121, 129, 130, 136, 143,
241, 258, 267, 268, 292, 296, 324, 325,
330-334, 344-346, 369, 394
Eisenhower, Dwight D.
(1890-1969): 295,299,310,312
Eliot, T.S. (1888-1965): 60,61,279,331,
366, 394
Ellis, H. Havelock (1859-1939): 342
Emerson, Ralph Waldo (1803-1882): 93
Engels, Friedrich (1820--1895): 260,277,
293,314,395
Engemann, Marguerite (see Margaret
Wiener)
Euclid (3651-275? B.C.): 326, 368
Euler, Leonhard (1707-1783): 62, 97,
345
Euripides (480?-406? B.C.): 370
Evans, W. 114
Fano, Robert (b. 1917): 373
Faraday, Michael (1791-1867): 24,99,
176
Faramelli, N.J. 333, 395
Ferry, D. K. (b. 1940): 173, 174
Feynman, R.P. (b. 1918): 84,128
Fisher, Sir Ronald A. (1890-1962): 153,
254, 395
Fisk, James B. (b. 1910): 312
Fock, V. (1898-1974): 121, 125,271
Foias, C. 193
Folin, Otto (1867-1934): 34
Forrester, Jay W. (b. 1914): 175
Forsyth, G. E. (b. 1917): 296
Foures, Y. 182,395
Fourier, Jean-Baptiste Baron de
(1768-1830): 97-99,104,144,163,177
Fowler, Henry Watson (1858-1933): 70
Francis of Assisi, Saint (1182-1226): 318
Frank, Philipp (1884-1966): 115, 136,
395
402
Frankel-Conrat, H.L. (b. 1910): 248
Frankfurter, Felix (1882-1965): 362
Franklin, Philip (1898-1965): 68,71
Frechet, Maurice (1878-1973): 74-77
Fredholm, I. (1866-1927): 163
Freudenthal, Hans (b. 1905): 349, 353,
355, 395
Frege, G. (1848-1925): 49-50,59
Freundlich, E.F. (1884-1964): 133
Freyman, M. 251
Friedrichs, K. (1901-1983): 101
Fuller. H.J. 280, 395
Gabor, Sir Dennis (1900-1979): 111
Gage, F.D. 393
Galileo (1564-1642): 315,368
Gallie, W.B. (b. 1912): 314,317,395
Galois, E. (1811-1832): 345
Garno, Hideya (b. 1924): 111
Gandhi, M.K., Matatma
(1869-1948): 318, 351, 395
Garcia Ramos, J. 197,205-208,248,
395
Gauss, Karl F. (1777-1855): 59,87
Gelfand, I.M. (b. 1913): 108
Getting, Ivan A. (b. 1912): 349
Gibbs, Josiah Willard (1839-1903): 55,
79, 122, 139, 143, 145, 146, 254, 354
Gilbreth, Frank (1868-1914): 254
Gilbreth, Lillian (1878-1972): 254
Giuculescu, A. 256, 395
Gleason, A.M. (b. 1921): 128,395
Glimcher, Melvin J. (b. 1925): 230
Glushkov, Y.M. (b. 1923): 260,261,395
Giidel, Kurt (1906-1978): 50,51,52,56,
136,219,220,261,346, 395, 396
Goethe, J.W. (1749-1832): 32,316,341
Golden, Reverend 370
Goldstine, H.H. (b. 1913): 175,232,
239, 396
Gordon, W. 120
Gould, K.E. 162,396
Goursat, E. (1858-1936): 353
Grattan-Guiness, I. (h. 1941): 43, 64,
396
Gray, T.S. (b. 1906): 162, 163,396
Green, G.F. (1793-1841): 87
Green, Gabriel Marcus (1891-1919): 77,
363
Greenberg, D. S. 361, 396
Hadamard, Jacques (1865-1963): 169,
170, 325, 396
Haldane, J.B.S. (1892-1964): 121-127,
136, 168, 248, 251, 254, 260, 293, 366,
396
Halmos, P.R. (b. 1914): 345-347,396
Hamilton, Sir William R.
(1805-1865): 142, 147, 345
Hantschius 127
Name Index
Hardy, G.H. (1877-1947): 22,47,55,
78, 94, 105-107, 135, 136, 335, 344,
345,353,371,396
Haug, E. 233
Hausdorff, F. (1868-1942): 84
Hazen, H.L. (1901-1980): 163,396
Healy, Sister E. T. 343, 393
Heaviside, Sir Oliver
(1850-1925): 100-103,336,340,346
Hegel, G.W. (1770-1831): 60
Heims, S. (b. 1926): 20, 242, 248, 293,
308,314,333,334,396
Heine, Heindrich (1797-1856): 335,341
Heins, A. E. 134
Heisenberg, Werner (1901-1976): 117,
118,121,184,241
Helmholtz von, H.L.F.
(1821-1894): 148
Heraclitus (6th-5th cent. B.C.): 268
Hermite, C. (1822-1905): 270
Heron of Alexandria (c. 100 A.D.): 256
Hershey, Alfred D. (b. 1908): 249
Hertz, H. (1857-1894): 24, 25
Hewitt, E. (b. 1920): 108, 114,396
Hilbert, D. (1862-1943): 26,59,60,94,
240, 371
Hill, G. W. (1838-1914): 207
Hille, Einar (1894-1980): 108
Hitchcock, Alfred (1899-1980): 94,
338-340
Hocking, William E. (1873-1966): 66
Hodgkin, Sir Alan Lloyd (b. 1914): 208
Homer (c. 850 B.C.): 293, 335
Hopf, Eberhard (1902-1983): 132-135,
139, 186, 196,396
Horace (65-8 B.c'): 335
Huntington, E.Y. (1874-1952): 43,46,
73-75, 89, 371
Hurewitz, Witho1d (1904-1956): 349
Husser!, Edmond (1859-1938): 59, 371
Huxley, A.F. (b. 1918): 208
Huygens, C. (1629-1695): 125
Ikehara, Shikao (l904?-1984): 107, 168,
375
Inge, W.R. (1860-1954): 369
Ingham, A. E. (1900-1967): 105
Ito, K. (b. 1915): 83,86
Jackson, Dougald C. (1865-1951): 96
Jacobs, W.W. (1863-1943): 316
Jacquard, J.M. (1752-1834): 219,220
James, William (1842-1910): 39,62
Jesperson, Otto (1860-1943): 280
Jewett, Frank (1879-1949): 18,356,360,
361
Johnson, Lyndon B. (1908-1973): 15,
364, 374
Jones, Rufus (1863-1948): 218
Name Index
Kac, Mark (1914~1984): 83,84,86,128,
396
Kahn, Henry 30
Kailath, T. (b. 1935): 135
Kakutani, S. (b. 1911): 88,130,151,
266,271,397
Kaluza, Th. (1885~1954): 119,121
Kant, Immanuel (1724-1804): 41,59,
71,72, 120,277,316,341
Karamata, J. 108
Kellogg, O.D. (1878~1932): 86,87,91,
361
Kelvin, Lord (Sir William Thompson)
(1824-1907): 102,103,161,254
Kennedy, Peggy (see Peggy Wiener)
Kepler, Johannes (1571~1630): 19,25,
345
Kerensky, A.F. (1881~1970): 32
Ketner, K.L. 221
Kettering, C.F. (l876~1958): 340
Keynes, Lord J.W. (1883~1946): 284,
397
Keyserling, Leon (b. 1908): 274
Khinchine, A.L. (1894-1959): 86,113,
149,194
Kipling, Rudyard (1865~1936): 335, 336
Kirchoff, G.R. (1824~1887): 176
Kleene, S. C. (b. 1909): 220
Klein, Felix (l849~ 1925): 225
Klein, O. 120, 121, 125
Kline, J. R. 167
Koebe, P. (l882~1945): 18
Kolmogorov, A.N. (1903~1987): 86,90,
149, 193~195, 241, 261, 397
Konoye, Prince (1891 ~ 1945): 298
Kosambi, D.D. (1907~1966): 363,397
Kronecker, L. (1823~1891): 59
Kropotkin, Prince P.L. (1842~lnl): 34
Lafargue, Paul (1842~1911): 292,397
Lagrange, J.L. (1736-1813): 87,99,345
Laguerre, E. (l834~ 1886): 164, 165, 270
Lambert, J.H. (l728~1777): 107
Landau, Edmund (1877~1938): 17,59,
60,371
Langmuir, Irving (1881~1957): 245
Laplace, P.S. (1748~1827): 87
Laski, Harold (l893~ 1950): 362
Lau, K-S. (b. 1948): 114,397
Lazarsfeld, P.F. (b. 1901): 290
Leahy, William D. (1875~1959): 296,
298
Lebesgue, H. (1875~1941): 77,78,83,
92, 139, 145, 147, 354
Lee, Yuk Wing (b. 1904): 165~168, 178,
187, 195,212,235,239,240,372,397
Legendre, A.M. (1752~1833): 163
Leibniz, G.W. (1646-1716): 19,39,76,
121,122,125,126,161,162,171,
219~222, 239, 254, 277, 319, 363, 366,
368
403
Lemaitre, Cannon G.H.
(l894~1966):
170
Lenard, P.E.A. (l862~1947): 315
Lenin, V.1. (1870~ 1924): 122, 260, 275,
397
Leontovich, M. 86
Leray, J. (b. 1906): 101
Lettvin, Jerome (b. 1920): 218
Levinson, Norman (l912?~1975): 92,
172, 349, 350, 353, 355, 375, 397
Levy, Hyman (l889~1975): 59,261
Levy, Paul (1886-1971): 86, 194
Lewis, Albert C. 18
Lewis, T. 204
Lichenstein, Leon (1878~1933): 136, 167
Liddell Hart, Sir Basil H.
(1895~1970):
298, 398
Lilienthal, D.E. (b. 1899): 311
Lindsley, D.B. (b. 1907): 236
Liouville, J. (l809~1882): 147
Littlewood, J.E. (l885~1977): 47,
105~1O7, 135, 136,368,371
L1ull, R. (1225~1316): 254
Lobatchevsky, N.1. (1793~ 1856): 73
Locke, John (l632~1704): 368
Lorente de No, R. (b. 1902): 239
Lovelace, Lady (Ada Byron)
(l815~?):
219
Lovell, C.A. 182, 189, 192
Lowell, Abbot Lawrence
(l856~ 1943):
32, 66, 362
Lucretius (9?~55 B.C.): 122
Lynd, Albert 280, 398
Mach, E. (1838~1916): 54,277
Mackey, G. W. (b. 1916): 128, 398
Maimonides, Rabbi Moses
(1135~1204):
41,43,398
Malliavan, P. 114
Mandelbrojt, Szolem (l899~ 1983): 170,
251
Mandelbrot, Benoit (b. 1924): 85,291,
368, 398
Mann, Robert W. (b. 1924): 230, 231
Marcinkiweicz, J. (1910-1940): 114
Marquand, A. 221
Marshall, George C. (1880-1959): 296
Martin, K. 362, 398
Martin, W. T. (b. 1911): 151,266,271,
393
Marx, Karl (1818~ 1883): 260, 277, 284,
292, 314, 322, 398
Masani, P.R. (b. 1919): 112,265,333,
398
Masaryk, Thomas (1850-1937): 31, 136
Maupertius, P.L. (l698~1759): 121
Maxwell, James Clerk
(1831~1879):
23~25, 81, 99,109,116,
143, 146, 155,254,258,334,345
McCulloch, Warren S. (l898~1969): 200,
205, 218, 219, 222, 223, 225, 227, 232,
404
236, 238, 239, 243, 248, 346, 366, 367,
398, 399
McMillan, Brockway (b. 1915): 151,375
McTaggart, 1.M.E. (1866-1925): 56,60,
371
Mehra,l. 118,398
Mendel, Gregor (1822-1844): 249
Menger, Karl (b. 1902): 136, 167
Mercer, J. (?-1932): 47
Meyer, Ella J. (Mrs. J. Franklin) 353
Michaelangelo, B. (1475-1564): 293, 347
Michaelson, A.A. (1852-1931): 110
Mill, John Stuart (1806-1873): 35
Millikan, Robert Andrew
(1868-1953): 366
Milton, lohn (1608-1674): 293
Milyukov, P.N. (1859-1943): 32
Minkowski, H. (1864-1909): 346
Mobius, A.F. (1790-1868): 63
Moliere (1. B. Poqug1in)
(1622-1673): 342
Monod, 1. (1910-1971): 249,398
Moore, E.H. (1862-1932): 77
Moore, G.E. (1873-1958): 56,371
Moore, O.K. 191
Moore, R.L. (1882-1974): 75
Mordell, L.l. (1888-1972): 47
More, Saint Thomas (1478 1535): 335
Morgenstern, O. (1902-1977): 268, 290,
296, 303,401
Morishima, M. (b. 1923): 314,399
Morris, William (1834-1896): 277
Morse, Marston (1892-1977): 108
Mozart, W.A. (1756-1791): 93,259
Muller, H.J. (1890-1967): 312
Mumford, Lewis (b. 1895): 284, 336,
337, 369, 399
Munsterberg, H. (1863-1916): 46
Nagy, B. Sz. (b. 1913): 193, 399
Napoleon, B. (1769-1821): 300,304
Nelson, Edward (b. 1932): 83, 128, 130,
241, 399
Nelson, Lord Horatio (1758-1805): 304
Neurath, O. (1882-1945): 336,399
Newton, Sir Isaac (1642-1727): 19,76,
87, 122, 342, 345
Nordhoff, C. (1830-1901): 29,399
Name Index
Paley, R.E.A.C. (1907-1933): 84,
136-138,140,166,169,177,197,372
Palmer, G.H. (1842-1913): 43,46,371
Parkinson, C. Northcote (b. 1909): 286,
287, 399
Pascal, B. (1623-1662): 19,161, 171,
239, 254
Pauling, L. (b. 1901): 312
Peano, Giuseppe (1858-1932): 53,59,
147
Pearl, Raymond 65
Peirce, Charles Sanders
(1839-1914): 15-17,22,39,55,62,74,
93,221,255,277,369,399
Perrin, J.B. (1870-1942): 79,81,82,84,
293, 366, 399
Perry, Ralph Barton (1876-1957): 43.
46,66
Phelan, G.B. (b. 1892): 367,399
Phillips, H.B. 71,88,171, 180
Phillips, Ralph S. (b, 1-913): 349,391
Pierce, J.R. (b. 1910): 155,257,399
Pincus, 1. 135
Pitt, Harry Ray (b. 1914): 108,372
Pitts, Walter (1923-1969): 201,207,208,
218,219,223-225,227,232,236,239,
243, 248, 398, 399
Planche rei, M. (1885-1967): 98, 147
Planck, Max (1858-1947): 117,130,156
Plato (427-347 B.C.): 15,24,43, 122,
124, 252, 253, 255, 259, 260, 297, 327,
335, 344, 363, 368
Poincare, H. (1854-1912): 59,62,207,
315
Poisson, S.D. (1781-1840): 87
Pollack, Rabbi Herman 370
P6lya, George (1887-1985): 47
Pope, Alexander (1688-1744): 342
Post, E.L. (1897-1954): 220
Potra, F. 232
Prokhorov, Yu. (b. 1929): 86
Pupin, M.1. (1858-1935): 336
Pythagoras (572-501 B.C.): 24, 345, 368
Quevedo, Tores y 304
Quine, W.V. (b. 1908): 52,54,55,399
Rabi, I. (b. 1894): 312
Raisbeck, Barbara (see Barbara Wiener)
Ramanujan, S. (1887-1920): 345,350
Randell, Brian 173, 175, 180, 399
Rashevsky, N. (b. 1899): 218
Rayleigh, Lord 1.W.S. (1842-1919): 102,
103, 111
Rechenberg, H. 118,398
Reuther, Walter (1907-1970): 16,62,
273-275,299,300,302, 303, 314
Riccia, 1. Della 128
Richards, LA. (b. 1893): 56
Rickover, Hyman (1900-1986): 281,299
Riemann, B. (1826-1866): 107,346
405
Name Index
Riesz, F. (1880- I 956): 75
Robinson, E.A. (b. 1930): 297
Roosevelt, Franklin D.
(1882- I 945): 296, 298
Rosenblith, W.A. (b. 1913): 235
Rosenblueth, Arturo (\ 900- I 970): 34,
197-205,207,208,210,21 1,218,225,
233, 239, 240, 243, 248, 249, 25 I, 36 I,
366, 367, 373, 399
Rosenthal, A. 147
Ross, K. A. (b. 1936): 108, I 14, 396
Rousseau, J.J. (1712-1778): 277
Royce, Josiah (1855-1915): 43,46,361,
371
Rudin, W. (b. 1921): 137,399
Runge, C. 180
Ruskin, John (\819-1900): 335
Russell, Bertrand (1872-1970): 18, 22,
45,47-52, 55, 56, 59, 60-64, 72, 79, 93,
199, 220, 239, 277, 297, 303, 335, 346,
366, 371,400, 402
Russell, H.N. (1877-1957): 243
Rutherford, Lord Ernest
(1871-1937): 123,322
Saeks, R.E. (b. 1941): 173,174,400
Samuel, A. L. (b. 1901): 306,400
Santayana, George (1863-1952): 43,61,
371
Santillana, George de 18
Sarvagatananda, Swami 370
Schetzen, M. (b. 1928): 271,400
Schickard, W. (1592-1635): 254
Schiller, J.C.F. (1759-1805): 341
Schmidt, K. 43, 46, 371
Schmidt, Robert 105-106
Schopenhauer, A. (1788-1860): 341
Schottky, W. 85
Schrodinger, E. (1887-1961): 120, 122,
123, 130, 241, 249, 345, 400
Schroeder, F. (1841-1902): 43,48
Schuster, Sir Arthur R.
(1851-1934): 102-104, 11 I-I 13,400
Schwartz, Laurent (b. 1915): 100, 251
Schweitzer, Albert (1875-1965): 318
Scimone, Frank J. 352-353
Segal, I.E. (b. 1918): 241,395
Shakespeare, William (1564- I 616): 347
Shannon, Claude E. (b. 1916): 50, I I I,
153-155, 159, 182,221,224,253,262,
356, 373, 400
Sheffer, Henry Morris (1883-1864): 55,
74
Siegel, Armand (b. 1914): 128, 129, 131
Simeone, S.A. 198
Sinclair, Upton (1878-1968): 362
Singleton, H. E. (b. 1916): 211
Smith, Adam (1723-1790): 281
Smoluchowski, M.V. 55,79,241
Snow, Sir C. P. (b. 1905): 359,400
Sobolev, S. L. (b. 1908): 101
51,295,
406
Vigier, J.P. 129
Vinci da, Leonardo (1452-1519): 18,343
Voltaire, F.M. (1694--1778): 122
Volterra, Vito (1860-1940): 77,163,264,
266, 271
Von Foerster, Heinz 248
Von Laue, Max (1879-1960): 111
Von Mises, R. (1883-1953): 86
Von Neumann, John (1903-1957): 50,
51-52,109,119,129,131,140,141,
184, 225, 232, 237, 239-242, 247, 248,
268,290,295,297,303,310-315,335,
345, 373, 396, 400, 401
Walsh, J. L. (1895-1973): 360
Walter, W. Grey (1910-1977): 212,234,
237,240,251,401
Walton, E. T. S. (b. 1903): 367
Watson, J.D. (b. 1928): 20,401
Watt, James (1736-1819): 252
Watts, A.W. (1915-1973): 25,401
Weaver, Warren (1894--1978): 183, 192,
207, 391
Weiss, P.R. 391
Wells, H.G. (1866-1946): 37,335
Weyl, Hermann (1885-1955): 53, 115,
120, 151,200,225, 345, 368, 369, 401
Whitehead, Alfred North
(1861-1947): 21,25,43,45,50,56,59,
72,93, 199,220,259,277,324,334,
346, 361, 367, 368, 369, 402
Whittaker, Sir Edmund
(1873-1956): 118,402
Wiener, Barbara (daughter)
(b. 1928): 94
Wiener, Bertha (sister) (b. 1902): 30, 35
Name Index
Wiener, Bertha Kahn (mother) 30, 371
Wiener, Constance (sister) (b. 1898): 30,
35
Wiener, Fritz (brother) (b. 1906): 30
Wiener, Leo (father)
(1862-1939): 29-47,67,94,335,361,
371
Wiener, Margaret (wife) (b. 1894): 38,
94, 95, 372
Wiener, Peggy (Margaret) (daughter)
(b. 1929): 94, 339
Wiener, Philip P. (b. 1905): 369
Wiener, Solomon (grandfather) 29
Wiesner, J. (b. 1915): 211,212,228
Wigner, E.P. (b. 1902): 367,402
Wilder, R.L. (b. 1896): 77
Wilhelm II, Emperor (1871-1918): 32
William of Ockham (c. 1200-1259): 62
Wilson, Woodrow (1856-1924): 67
Witner, A. (1903-1958): 141, 151,204
Wittgenstein, Ludwig (1889-1951): 52,
53, 57,402
Wold, H. 193
Wolf, J. 111, 393
Wolfson, Harry A. (1887-1974): 69
Womack, J. Printise 275-276
Wrinch, Dorothy 245
Yaglom, A.M. 113,193,401
Young, L. C. (b. 1905): 136
Young, R. C. 136
Young, W.H. (1863-1942): 136
Zangwill, Israel (1864--1926): 34
Zermelo, E. (1871-1953): 51
Zygmund, A. (b. 1900): 84
407
Subject Index
Abduction 62
Academie Francaise 358, 359
Accademia dei Lincei 358
Active and passive mechanisms 202
Albany, New York 67
Algorithm 219,220
Alice in Wonderland 93, 152
American Mathematical Society
1934 Colloquium Lectures 137, 372
1940 Dartmouth meeting 172
1949 J.W. Gibbs Lecture 373
American Society of Mechanical
Engineers 277
Analysis situs 63
Angelos 251
Animal feedback (homeostatic, postural,
voluntary) 210-211
Anti-aircraft fire control 181
cranking errors 189
error of performance 186, 187
predictor 187
transient errors 189
Arabian Nights 316
Art
Dante's view of 343
Eckhart's view of 343
presentative aspect of 344
representative aspect of 344
Arthasastra 287
Artist and Artisan 343, 346
Atomic diplomacy 295, 311
Atomic energy
and Oakridge National
Laboratory 312
and Price-Anderson Act 312
Atomic warfare
atomic bomb 301,310
atomic bombing of Japan 296, 298
civil defense 301
Litianthal Report 311
Attica prison 351-353
Automata theory
game-playing automata 304, 305, 306,
355-356
408
Bush integraph 160, 161
Bush network analyzer 160
Calculi (logical) 54
Calculus of variations 186
Calculus ratiocinator 219, 221
Cambridge University 45,66, 122, 136,
372
Causal operators 101
Causality and analyticity 101, 134-135
Cell, complexity of 244-245
Chaotic integrals 265
Chess-playing automata 304-305,
355-356
Child training, Leo Wiener's views
on 35
Clonus (see Muscle clonus)
Coherence of light 110--111
Coherency matrix 110--111
Cold War 313
Columbia Medical School 218
Columbia University 61, 371
Communication 152 (see also
Information)
and time scales 307-308
as social cement 288
channels of 288-289
communal information and
cohesiveness 288
communications engineering 101-102,
206
Communism 332
Communists 293, 294
Complementarity 116-117
Complex number field C 89
Computer (see also Turing machine)
abacus 171
all-purpose 219
analogue 110, 160-161
as prosthesis for the brain 232, 233
data-storage by tape 172, 173, 174
digital 171
electronic 15, 172, 224
ENIAC 174, 239, 244
exchangeability of time and
space 226-227
necessity for high speed 171-172, 173
program 220, 238
special purpose 171
special purpose digital 219
solution of PDE's by scanning
179-180
transistorized 174
use of vacuum tubes in 172, 174
Computing machines (see Computer)
Conditional Banach spaces 114
Consonance 253
Contest 268
Control (see Anti-aircraft fire control;
Automatic control; Homeostasis;
Animal feedback)
Subject Index
Controlled experiment 152
Convex bodies 72
Convolution filters 10 I, 107, 163-164,
177-179
Convolution integrals 101, 106, 107, 264
Copyright 338
Curl 23
Cybernetics (defined) 256
cybernetical attitude 239-240, 253-256
Cybernetical Circle 239
origin of idea of 56, 57, 58
origin of name of 251-252
Soviet attitudes toward 260-261
Deterministic dynamical system 141-142
Diagonalization procedure 48
Dialectic duality 25-26
Dialetic materialism 293
Dirichlet problem 87-88, 171, 180
Docent Lectures 62-64
Dual nature of light 25-26, 116
Duty 66, 125, 327
Dynamical system
coefficients of inertia 157
conservative 142
deterministic 141-142
Hamiltonian 142, 157
Hamilton's canonical equations 142
kinetic energy 157
phase space 142
potential (energy) 87, 157
state space 142
Econometrics 291
Economic planning 285-286, 329
Economic system (U.S.) 281-283
acquisitiveness in 283, 284
alienation in 284
and hard-work ethic 283
credit-card hedonism in 283-284
entrepreneurship in 282
investments in 282-283
irreligious aspect of 284
market as n-person game in 281
promotion in 283
volatility of 283
Education 277-281
and Parent Teachers Association
(PTA) 281
and teachers' colleges 280
and teachers' trade unions 281
damage from harshness in 35
decline of in U.S. 281
John Dewey's ideas on 277-280
quackery in public schools 280
Electric engineering 10 I-I 02
Electric filters
amplitude response function 165
causality (physical realizability) 101,
138, 163-164
409
Subject Index
frequency response function 164, 165
linearity 101
lumped-passive 164, 176
phase-response function 165
weighting function 163
Electric networks (see Electric filters)
Electrodynamometer 160
Electroencephalogram (EEG) 233 (see
also Brain-wave encephalography)
Electromagnetic theory of light 24, 109
Electron microscopy 245, 246
Electronic computer 15, 172, 224
Electronic Age 25
Encvclopedia Americana 67 372
'
Energy 87, 148, 157
ENIAC 174, 239, 244
Entrainment of frequency 237
Entropy (see Thermodynamics)
Ergodic hypotheses 146-147
Ethics 58, 125, 311, 327
altruism 125, 327
conscience 58
duty 66, 125, 327, 328
vocation 327
Evil
analogy of, with entropy 319
as corruption 319 (see also Fall of
man)
moral 319, 322
natural 319
non-eradicability of 319
phylogenetic origins of 319, 320
St. Augustine's views on 318,319 321
Explication 72
'
Extensive abstraction 56
non-differentiable 79, 84
periodic 97
Functional integration 84
Gadget-worship 171
Game theory 268-269, 286, 287, 290,
303, 306, 309
Gating mechanism 236
Gauge theory 120
Gene- reproduction 124
General Electric 67
Generalized Harmonic Analysis
(GHA) 102, 104--105, 184
almost periodic functions 104, 141
(auto-)covariance (or correlation)
function 102, 110, 113, 186
cross-covariance 110
Einstein's contributions to 112-113
Generalized Bessel identity 104--105
generalized Fourier coefficients 141
generalized Fourier transform 104
periodogram 104
spectral distribution and densitv 104,
112, 113
'
total mean-power 109
Geometry
Einstein's views on 53, 72
Euclidean 73, 342
Greek conception of 73
Kant's views on 71-72
non-Euclidean 73
German intellectual tradition 32-33
German romantic movement 341, 342
Gestalt (see Pattern)
Glasnost 293
Gleason's theorem 128
God (Logos) 25, 292, 330, 332-334
Jehovah 333
of Spinoza 25, 292, 333, 345
personal 333
Giidel number 346
Giidel incompleteness
metatheorem 52-53
Gorgias 252
Giittingen University 56, 59-60, 105,
117, 371
Gravitational force 23
Guggenheim Fellowship
Gulliver's Travels 335
105, 372
410
pure tone 96
sin usoidal curve 97
Harmonic analyzer 161
Harvard Lampoon 362
Harvard reserve regiment 66--67, 371
Harvard University 39-41, 43-44,
63-64, 361-362
Hausdorff dimension 83, 85, 90, 91
Heart, human 204
auricles 204
dis tole 204
fibrillation 204
flutter 204
sistole 204
tonic-clonic cortical
responses 204-205
Heat equation 84
Heaviside operational calculus 100, 160
Hebraiche Melodien 335
Helices (spirals) 108, 114, 195
Hidden parameters 129
Hidden periodicities 104, 112-113
Highest good 58
Hilbert space 26, 108, 114, 118-119,
128-129, 131, 193, 195,241
Hilbert transform 165-166
History, episodic view of 297-298, 302
Hitopedesa 287
Holography III
Homeostasis 34
breakdown of 331
homeostatic diseases 211
homeostatic feedback 210
Homeostat 212
Hominids 322, 323
Homogeneous chaos 149
Homogeneous random fields 141
Homo peccator 322
Homo sapiens 322
Hopf-Wiener integral equation 133-134,
186,241
Hunting 201
Idealization 72, 116, 122, 184,258-259,
326
Incoherent signals 110-111
Indeterminism 146, 253, 254-255, 290,
291,318 (see also Uncertainty
principle)
Indian Statistical Institute 193, 285, 351,
373
Industrial revolutions (First, Second,
Third) 249,272,273,281,346
Infinite (actual) 59
Information theory 153-156
communal 288
informative value of a
message 153-154
loss of 228, 229
misuse of 288-289
optical III, 228
Subject Index
Inquiry 268
Instruments 186, 187
limitations of 143, 146
response time (or latency) 109
Instrumentation theory 186, 187, 195,
196
Integraph 160,161,162,175-176
International Copyright Convention 337
International Mathematical Congresses
1920 Strasbourg 75, 372
1932 Zurich 136, 166,372
1936 Oslo 170,231,372
1950 Cambridge, Mass. 373
Iron lung 229
Isomorphism 200
Jacquard loom 219
Japan, atomic bombing of 296, 298, 323
Jewish Numeras Clausus 32,361-362,
363
Jewish question 30-32,41-43, 59
John Thornton Kirkland Fellow 44, 371
Josiah Macy Foundation 205, 373
Judaism 42
Juniata College 94
Kinesthetic organs (proprioceptors)
Klein-Gordon equation 120
Kolmogorov-Wiener prediction
theory 193
Korean War 273, 299, 300
213
Labor, American
migrant 275-277
retraining of 277
Lag mechanisms 19 I
Laguerre functions 164-165, 177
Laplace's equation 87
Latency 191
Lead mechanisms 191
Learning 152-153,262-264 (see also
Education)
Learning mechanisms (machines) 58,
262
Lebesgue integral and measure 78, 83,
99, 104, 140
Lee-Wiener network 163-165,177-179,
195, 264, 270
Leibnizian philosophy 121-122,
125-128,219
calculus ratiocinator 219,220
machina ratiocinatrix 219, 220
monads 121, 125-126
principle of identity of
indiscernibles 121
principle of sufficient reason 121-122,
126
Leukemia 211
Libertinism 47
Linear operator 99-100
Linear partial differential equations 99
Subject Index
53, 54
Machine 256
Machine intelligence, creativity,
tropism 212, 346--348 (see also
Computers)
Homeostat (machina sopora) 212
machina ratiocinatrix 219-222
machina speculatrix ("tortoise") 212
"moth-bedbug" 211
reproducing machines (see Selfreproducing mechanism)
Machine-man concatenation 188-189,
310
"Mad Tea Party" of Trinity 56
Manhattan Project 310
Marxism 292, 293, 294
Massachusetts General Hospital 234
Massachusetts Institute of Technology
(MIT) 71,92,96, 181,207,218,234,
372
MIT Autocorrelator 234, 235
MIT Radiation Laboratory 349
411
Molecular biology 248-250
Monads 121, 125-126 (see also
Leibnizian philosophy)
Monkey's Paw, The 316
Monogenism 321,323-324
Monte Carlo method 151
Multivariate prediction 193, 297
Muscle clonus 205-207
Music, musical notation 116
Mutualistic Societies in America 29-30
Myopia 37
Myths 324-325
Mythic mode of perception 25, 324-326
Mythology 325-326
Name and object 325
National Academy of Sciences
(U.S.) 356--358,360-361,373
National Defense Research Committee
(NDRC) 181-182,373
Project DJ.C. 5980 182
National Institute of Cardiology,
Mexico 197,207
National Medical of Science 15,364
National University of Mexico 373
Negentropy 155
Nerves 215
Nervous system, human 213,243
action potential 213
afferen t nerves 215
autonomic 215
center of nerve 215
central 202,213
efferent nerves 215
gating 236
gray matter 213
mixed nerves 215
parasympathetic 215
peripheral 214
physiological clock 236
reflex action 208
spatial summation 214
sympathetic 215
synapse 213
synaptic delay 214
synaptic excitation 208
temporal summation 214
white matter 213
Neuron 213
association 214
axons 207, 213
cytoplasm 213
dendrites 213
excitatory 214
firing of 213,214
inhibitory 214
irritability of 213
motor 214
nucleus of 213
quiet 214
refractory period of 214
412
resting state of axon 2 \3
sensory 214
Neurophysiology 249, 250
analogical aspects of 244, 262-263
New England (democracy of) 34, 38,
289
New York Times 15,229
Newtonian mechanics (see Dynamical
system)
Noise 19,143
Non-linear filters (networks,
transducers) 264-267,270--271
Nucleic acid complexes 249
Observation, photometic nature of 119
Observer-observed coupling 127-128,
289, 290, 291
On a Balcony 30
Ontogenetic learning 262-263
Operationalism 121
Operator
linear 99-100
non-linear 264
Optics 109-112
brightness 109
coherence 110-111
Michelson interferometer 110
photometer 109
polarized light III
Optical depth 132
Ordered pair (couple) 55
Organization 96
Paley-Wiener criterion 138
Paradise Lost 93
Paradoxes 48-50, 52
Parkinsonianism 210
Partial differential equations 99
parobolic 174
Particle-wave dichotomy 26, 27, 121
Pa th integral 128
Pattern (Gestalt) 96, 244
as group-invariant 225
Pattern recognition 15, 225-226
apparition 225
limitations of theory of 244
Peabody School 35,37,371
Peano's postulates 53, 65, 73
Perception
logical mode of 25
mythic mode of 25, 325
sensory 199-200
Philosophies of mathematics 54
Philosophy of Science Club 197
Photometer 109
Photosynthesis 156
Phylogenetic corruption 321,323-324
Phylogenetic learning 263-264
Physiological clock 236-237
Planning 285, 329
Platonism 258-260
Subject Index
Plattsburg, New York 66
Poetry 25
Polarization of light III
Polynomial chaos 150
Postulate systems 73-79
Potential theory 86-88
capacity 88
Dirichlet problem 87-88
the potential 87
Potentiometer 160
Power engineering 102
Pragmaticism 62
Pragmatism 39, 57, 62
Prediction theory 193
Prime number theorem 107
Principia Mathematica 45, 48, 50, 52, 53,
54, 55, 93, 220, 240
Principle of least action 121
Probability distribution 86
Probability theory 86
Project D.LC. 5980 182
Prometheus 325, 332
Proprioceptors (kinesthetic organs) 213
Prosthesis (muscular-skeletal and
sensory) 227-231
Boston Arm 229-231
Iron lung 229
sound communication with the
deaf 228-229
Protein molecule 124
Psychic signs 325
Pure tone 96, 115
Purpose (see Teleology)
Purpose tremor 20 I
Purposive action 124
Quantum mechanics 26,115-131
degeneracy 123
hidden parameters 129, 131
potential barrier 123
pure state 128, 131
role of Brownian motion in 128-131
self-repair in 123
state 128
Quasi-analytic functions 169
Quasi-ergodic hypothesis 147
Quickening 191
Radiative equilibrium 132
Ramakrishna Vedanta Society of
Boston 370
Recursion 50, 53
Reflex action 208-209
Refugee scientists 167-168
Relation structure 199, 200
Relation theory 50, 55, 63-64
Relativism 57, 58, 60
Relativity theory 26, 120
Religion (see also Evil; Fall of man; God)
"acts of God" 330
"acts of Grace" 330, 332
413
Subject Index
Adam and Eve 321
and science 324, 330, 333-334
Black Mass 315-316
Grace 321
role of death 319
self-alienation 322
simony 316
sin 322
Republic 43, 67
Retrospective operator 101
Riemann zeta function 107
Ritual 324
Robot arm 232-233
Rockefeller Foundation 205, 207
Royal Society 358, 359
Russell's antinomy ("paradox") 48-50,
52
Sage School of Philosophy 40, 41, 371
Scannmg 171, 173, 179-180,226-227
Schwartz distribution 100
Science fiction 335-336
Science and human welfare 285,291,
292,311,314-315,329-330
Scientific methodology
centrality of relation-structure 198,
199,200
complementarity 116-117,127-128
cybernetical augmentation of 253-255,
257
explication 72
faith as a part of 24-25, 326
function of science 358
idealization as a part of 258-260, 326
incomplete orderliness of the
world 146
instrumental limitation 143, 146
logical empiricist theses 53-54
mythic perception as a part of 25,
324-326
nature of inquiry 268-269
nature of logic 53-54, 222-223
observer-observed coupling 127, 289,
290, 291
organization of empirical data 41, 72
place of electronic computer in 232
233
'
propaedeutic role of mathematics
in 22-28, 367, 368
Secrecy 191, 192, 299
Self-alienation 322
Self-organizing systems 236, 237, 246
Self-reproducing mechanisms 246,
263-264, 266-267, 269
Semi-exact sciences 290
Sensory prosthesis (see Prosthesis)
Servomechanisms 189,202,210
Set, concept of 48
Shannon-Wiener information theory (see
Information theory)
Sheffer's connective 74
Shot effect 85
Signal 19, 102 (see also Information
theory)
Signal detection 257
Simony 316
Sin 322 (see also Evil)
Sinusoidal curve 97
Social cohesiveness 289
Society for the Advancement of
Management 276, 277
Sociology, observer-observed coupling
in 289
Sorcerer's Apprentice 316
Sorcery 316
South Tamworth, New Hampshire 38
Soviet attitudes 229,251,260-261,314
Space-time continuum 26
Spectral distribution and density 104,
112, 113
Spectral synthesis 108, 114
Spectral theory of operators 193
Spike potential 207,215-217
differential equation of 217
Hermann model 216
Spinal cord 208, 209
Spinoza's God 25, 292, 333, 345
Spirals (helices) 108, 114, 195
State, the 328, 329
Stationary random measure 149 195
Stationary sequences 193, 195 '
Statistical mechanics
and communications engineering 102,
154-155
Boltzmann's H-theorem 149
complexion of a gas 159
ergodic hypothesis 146
ergodic theorem for homogeneous
chaos 150
Maxwell's demon 155-156 325
microstate of a gas 158 '
quasi-ergodic hypothesis 147
statistical entropy 148
Statistical thermodynamics 148
Stochastic integration 84, 150
Stokes's theorem 23
Stratagem 267-269, 286-287, 303-306,
307,309,316-317
Strategic evaluations 306
Strategic learning 268
Switching networks 221
Symbolism, mathematical 21-28
Synthetic a priori 71,72
Synthetic logic 63
Synapse 213
Tala Institute of Fundamental
Research 373
Tauberian theory 105-108
Tchestnost 293
Teleological Society 239
414
Teleology 57, 124, 152-153,201-202,
253, 368
purposive mechanism 202, 262, 268
teleological mechanism 202, 226
Television scanning 171, 173
Templer, The 336
Tensor 26
Terry Lectures 306, 331, 374
Theory of messages 251 (see also
Message)
Thermodynamics
absolute temperature 148, 157
adiabatic transformation 147, 157
Caratheodory's principle 157
empirical temperature 157
entropy 147, 157
Helmholtz free energy 148
internal energy 157
state space J47
Time
anistropic 57. 152
Bergsonian 58, 152
Newtonian 58
Time Series
Topological transformation 75
Topology 62-63,74-75
Traclatus logico-philosphicus 53, 54
Trade union consciousness 275
Tragedy and catharsis 331-332
Transducers 256, 308
Transfinite induction 48
Transformation (see Operator)
Tsing Hua University 168, 372
Tufts College 39, 371
Turbulence 79, 151
Turing machine 173,219,220,221,222,
223, 238, 346, 348
Type theory 47, 50-52
Uncertainty principle 115-117, 119, 129
Unified field theory 120
United Automobile Workers of America
(UA W) 273-275
Subject Index
Universal Turing machine (see Turing
machine)
University of London 94
University of Maine 65
University of Melbourne 94
University of Missouri 30
University of Nancago 251
USSR 313,314 (see also Soviet
attitudes)
Values (see Ethics)
Vector 23, 75-76
Vector space (metric, normed)
Virus 244-245, 248
Vocation 327
Vocoder 228
Voluntary activity 188-189
75-76,78
182
415
416
Vannevar Bush (p. 166). Courtesy of The
MIT Museum (Photo VB 2).
Y. W. Lee (p. 167). Courtesy of The MIT
Museum (Photo YWL 2).
Wiener with colleagues (p. 169). Courtesy of
The MIT Museum (Photo NW 67).
Julian Bigelow (p. 183). Courtesy of Mr.
Julian Bigelow.
Academician Andre Kolmogorov (p. 194).
Courtesy of Birkhiiuser Boston Inc.
Arturo Rosenblueth (p. 199). Courtesy of
Mrs. Virginia Rosenblueth.
Norbert Wiener with Drs. J. Weisner and
Y.K. Lee (p. 212). Courtesy of The MIT
Museum (Photo NW 26).
Warren McCulloch (p.222). Courtesy of
The MIT Museum (Photo WSM 4).
Walter Pills (p. 224). Courtesy of The MIT
Museum (Photo WHP I).
The Boston Arm (p. 230). Courtesy of Mr.
T. Walley Williams, Liberty Mutual Insurance Company, Boston.
The Boston Elbow (p. 231). Courtesy of Mr.
T. Walley Williams, Liberty Mutual Insurance Company, Boston.
W. Grey Walter (p. 235). Courtesy of Burden Neurological Institute.
John von Neumann (p. 247). Courtesy of the
Niels Bohr Institut.
Henri Bergson (p. 255). Reprinted from:
Willy Haas: Nobelpreistrager der Literatur. Ein Kapitel Weltliteratur des zwan-
Photograph Index
zigsten Jahrhunderts. Heinz Moos Verlag, Heidelberg 1966.
William Ross Ashby (p. 259). Courtesy of
Professor N. N. Rao, Electrical & Computer Engineering Department, University of Illinois.
Walter Reuther (p. 274). Courtesy of Mrs.
Norma Bogovich, UA W Public Relations and Publications Department Detroit, MI. Photo by Alexander Archer,
UAW Solidarity Magazine.
John Dewey (p. 278). Reprinted from:
Living Philosophers, New York 1931,
p. 21. Courtesy of Simon & Schuster.
Chess playing automaton, built by Tores y
Quevedo in 1890: (p.304). Courtesy of
The MIT Museum (Photo NW 55).
Chess pla,ring automaton, the Belle (p. 305).
Courtesy of Dr. Kenneth Thompson,
Belle Telephone Laboraties.
St. Augustine in His Stu,~v (p. 320). Courtesy of Scala Istituto Fo\ografico Editoriale S.p.A., Firenze.
Homo peccator (p. 323). Courtesy of Ringier Dokumentationszentrum Ziirich.
(Photo WK 2 Japan II Dukas, Hiroshima (G 0136232)).
Norbert Wiener in the Soviet Union (p. 364).
Courtesy of The MIT Museum (Photo
NW 14).
Norbert Wiener at the White House (p. 364).
Courtesy of The MIT Museum (Photo
NW 51).